Loading models
381 models from 48 providers. Same prompts, real outputs, blind-vote results.
381 models
OpenAI
OpenAI GPT-6.1 Sol updates GPT-6 Sol for coding, computer use and document work. It accepts text, images and files, with a 1.05M-token context window and up to 128K output tokens. Five reasoning effort levels run from low through max.
Conversation · Reasoning · Code generation +4
Released 2026-09-29Large model
Anthropic
Anthropic's direct upgrade to Sonnet 5 for everyday coding and knowledge work. Anthropic says it writes output more than 30% faster than Sonnet 5 and finishes tasks with fewer tokens and tool calls, at the same price and a 1M-token context, with text, image and file input.
Conversation · Reasoning · Code generation +3
Released 2026-09-28Large model
Recraft
The speed and cost tier of Recraft V4.1 for text-to-image generation. Produces raster images at roughly 1K resolution across several aspect ratios, priced at $0.007 per image on the listed OpenRouter endpoint.
Image generation
Released 2026-09-23Large model
OpenAI
OpenAI GPT-6.1 Sol updates GPT-6 Sol for coding, computer use and document work. It accepts text, images and files, with a 1.05M-token context window and up to 128K output tokens. Five reasoning effort levels run from low through max.
Conversation · Reasoning · Code generation +4
Released 2026-09-29Large model
OpenAI
GPT-6 Sol sits between Astra and Luna in OpenAI's GPT-6 family. It accepts text, images, and files, with a 1.05-million-token context for professional tasks, coding, and reasoning.
Conversation · Reasoning · Analysis +2
Released 2026-09-22Large model
OpenAI
The fast, lower-cost tier of GPT-6 for chat, classification, and lightweight agent workflows. Luna accepts text, images, and files with the same 1.05-million-token context as Sol.
Conversation · Reasoning · Analysis +2
Released 2026-09-22Large model
OpenAI
GPT-6 Sol served with pro reasoning mode for more demanding tasks. It uses the same underlying model and 1.05-million-token context as Sol, with additional reasoning computation.
Conversation · Reasoning · Analysis +2
Released 2026-09-22Large model
OpenAI
GPT-6 Luna served with pro reasoning mode. It applies additional reasoning computation to the Luna model while retaining its text, image, and file inputs and 1.05-million-token context.
Conversation · Reasoning · Analysis +2
Released 2026-09-22Large model
OpenAI
The precision tier of OpenAI's GPT Image 2.5 pair and its most capable image model so far. Higher image quality than GPT Image 2, sharper detail, stronger style adherence and tighter control over edits. Built for campaign creative and polished product imagery.
Image generation
Released 2026-09-08Large model
OpenAI
The speed tier of OpenAI's GPT Image 2.5 pair. Image quality on par with GPT Image 2 at about half the latency, with the new xhigh and max quality presets, custom sizes up to 4K and tighter subject preservation on edits. OpenAI's default recommendation for everyday API image generation.
Image generation
Released 2026-09-08Medium model
OpenAI
GPT-6 Astra served with reasoning mode set to pro, which spends a larger thinking budget per request. Same underlying model as GPT-6 Astra, same 1.05M context and 128K output ceiling, aimed at the harder end of coding, analysis and long multi-step work.
Conversation · Reasoning · Code generation +4
Released 2026-09-04Extra large model
OpenAI
OpenAI's first GPT-6 model, out September 3 2026 from what the company calls its largest training run so far. It takes 1.05M tokens of context, writes up to 128K, and exposes five reasoning effort levels from low through max. OpenAI aims it at computer use, browsing, software engineering and long multi-step work.
Conversation · Reasoning · Code generation +4
Released 2026-09-03Extra large model
OpenAI
GPT-5.6 Luna is a fast, cost-efficient model in OpenAI's GPT-5.6 series, suited for high-volume, latency-sensitive tasks such as chat, classification, and lightweight agentic workflows, with capable reasoning for its price tier.
Conversation · Reasoning · Code generation +2
Released 2026-07-09Medium model
OpenAI
The middle tier of OpenAI's GPT-5.6 series, between Sol and Luna. For everyday coding, reasoning and agent work, at roughly half the cost of Sol.
Conversation · Reasoning · Code generation +2
Released 2026-07-09Large model
OpenAI
The top tier of OpenAI's GPT-5.6 series, for complex reasoning, coding and agent workflows. Strongest at command-line and multi-step coding, and at long-horizon problems.
Conversation · Reasoning · Code generation +3
Released 2026-07-09Extra large model
OpenAI
GPT-5.6 Luna Pro is the same underlying model as GPT-5.6 Luna, served with reasoning mode set to pro for higher-quality responses on complex tasks.
Conversation · Reasoning · Code generation +2
Released 2026-07-09Medium model
OpenAI
OpenAI's GPT-5.5, for complex professional work. Builds on GPT-5.4 with stronger reasoning, higher reliability and better token efficiency on hard tasks. 1M+ token context window (922K input, 128K output), text and image in.
Conversation · Reasoning · Code generation +2
Released 2026-04-24Extra large model
OpenAI
OpenAI's second-generation image model. Near-perfect typography (including non-Latin scripts), 2x faster than gpt-image-1.5, up to 2K resolution, and 'thinking' over the prompt for dense compositions.
Image generation
Released 2026-04-21Large model
OpenAI
GPT-5.4 at a smaller size, for high-throughput work. Text and image input, with reasoning, coding and tool use intact at lower latency and cost. The intended fit is chat products, coding assistants and agent workflows running at scale.
Conversation · Reasoning · Code generation +1
Released 2026-03-17Medium model
OpenAI
The smallest and cheapest model in the GPT-5.4 family, for speed-critical, high-volume work. Text and image input. It trades reasoning depth for responsiveness, which suits classification, data extraction, ranking and sub-agent calls.
Conversation · Reasoning · Code generation +1
Released 2026-03-17Small model
OpenAI
OpenAI's GPT-5.4, folding the Codex and GPT lines into one system. Over 1M tokens of context, 922K in and 128K out, with text and image input. The improvements land in coding, document understanding, tool use and instruction following.
Conversation · Reasoning · Code generation +2
Released 2026-03-05Extra large model
OpenAI
The Pro tier of GPT-5.4, with more reasoning compute for high-stakes work. Same unified Codex and GPT architecture, over 1M tokens of context at 922K in and 128K out, text and image input. Aimed at agentic coding, long-context workflows and multi-step problems.
Conversation · Reasoning · Code generation +2
Released 2026-03-05Extra large model
OpenAI
An update to the model most ChatGPT conversations run on. It answers with better context, refuses less often, and drops the caveats and hedging that used to interrupt a conversation. More willing to engage with nuanced topics, its prose reads cleaner, and reasoning, coding and creative work hold up.
Conversation · Reasoning · Code generation +1
Released 2026-03-03Large model
OpenAI
OpenAI's agentic coding model at the 5.3 mark, pairing GPT-5.2-Codex software engineering with GPT-5.2's broader knowledge. Leads SWE-Bench Pro and scores well on Terminal-Bench 2.0 and OSWorld-Verified. Built for long tool-using runs, and steerable mid-execution.
Conversation · Reasoning · Code generation +2
Released 2026-02-24Extra large model
OpenAI
OpenAI's latest image generation model with strong instruction following, optional transparent backgrounds, and quality controls.
Image generation
Released 2025-12-16Large model
OpenAI
GPT Image 1.5 with `quality=low` for faster and cheaper generations.
Image generation
Released 2025-12-16Large model
OpenAI
GPT Image 1.5 with `quality=medium` for balanced cost and quality.
Image generation
Released 2025-12-16Large model
OpenAI
GPT Image 1.5 with `quality=high` for maximum fidelity.
Image generation
Released 2025-12-16Large model
OpenAI
The fast, lightweight member of the GPT-5.2 family, built for low-latency chat. Adaptive reasoning lets it think only on the harder queries, so mathematics, coding and multi-step questions get more accurate without slowing ordinary conversation. Warmer and more conversational by default than the base model.
Conversation · Reasoning · Code generation +1
Released 2025-12-10Large model
OpenAI
The GPT-5.2 base model, with better agentic and long-context behaviour than GPT-5.1. Adaptive reasoning allocates compute per query, answering simple things quickly and spending depth on hard ones. Gains show up across mathematics, coding, science and tool calling, with more coherent long-form answers.
Conversation · Reasoning · Code generation +1
Released 2025-12-10Extra large model
OpenAI
The Pro tier of GPT-5.2, aimed at agentic coding and long-context work beyond what GPT-5 Pro handled. Test-time routing and intent cues like "think hard about this" carry over, with lower hallucination and sycophancy than the previous Pro.
Conversation · Reasoning · Code generation +1
Released 2025-12-10Extra large model
OpenAI
The largest Codex tier in OpenAI's GPT-5.1 line, built for coding agents that run for hours rather than turns. Training data stops in September 2024.
Conversation · Reasoning · Code generation +1
Released 2025-12-04Extra large model
OpenAI
The latest of the GPT-5 series: better general reasoning, tighter instruction adherence and a more natural conversational style than GPT-5. Allocates compute adaptively, answering simple queries fast and spending longer on hard ones.
Conversation · Reasoning · Code generation +1
Released 2025-11-13Extra large model
OpenAI
GPT-5.1 Chat (Instant), the fast and light member of the 5.1 family, built for low-latency chat. It still thinks selectively on harder queries, which keeps math, coding and multi-step accuracy up without slowing normal conversation.
Conversation · Reasoning · Code generation +1
Released 2025-11-13Large model
OpenAI
A smaller, faster GPT-5.1-Codex for coding at lower latency.
Conversation · Reasoning · Code generation +1
Released 2025-11-13Extra large model
OpenAI
GPT-5.1 tuned for software engineering: interactive development sessions and long independent runs, covering projects from scratch, feature work, debugging, large refactors and code review.
Conversation · Reasoning · Code generation +1
Released 2025-11-13Extra large model
OpenAI
The Pro tier of GPT-5, given more test-time compute for complex, high-stakes tasks. Same routing controls and intent cues as GPT-5, with lower hallucination and sycophancy and stronger coding, writing and health answers.
Conversation · Reasoning · Code generation +1
Released 2025-10-06Extra large model
OpenAI
GPT-5 retuned for Codex, OpenAI's agentic coding tool. It runs on the Responses API only, not chat completions, and its training data stops in September 2024.
Conversation · Reasoning · Code generation +1
Released 2025-09-23Extra large model
OpenAI
OpenAI's GPT-5, built around step-by-step reasoning and instruction following on high-stakes work. It routes at test time and reads intent cues like "think hard about this". Hallucination and sycophancy are down against GPT-4o, with better coding, writing and health answers.
Conversation · Reasoning · Code generation +5
Released 2025-08-07Extra large model
OpenAI
A compact GPT-5 for lighter reasoning, with the same instruction following and safety tuning at lower latency and cost. It is the successor to OpenAI's o4-mini.
Conversation · Reasoning · Code generation +1
Released 2025-08-07Medium model
OpenAI
The smallest model in OpenAI's GPT-5 line, priced for high-volume classification, extraction and short code edits rather than hard reasoning. Text and image input, training data through May 2024.
Conversation · Reasoning · Code generation +1
Released 2025-08-07Small model
OpenAI
OpenAI's smaller open-weight release, 21B parameters on a mixture-of-experts design. Matches o3-mini on quality while fitting in 16GB of memory, which puts it on consumer hardware and edge devices.
Conversation · Reasoning · Code generation +4
Released 2025-08-05Large model
OpenAI
OpenAI's first open-weight model since GPT-2, 117B parameters on a mixture-of-experts design. Reasoning depth is configurable per request and tool use is native, aimed at agent and production workloads.
Conversation · Reasoning · Code generation +3
Released 2025-08-05Extra large model
OpenAI
OpenAI's first generally available audio model. Features an upgraded decoder for more natural sounding voices and maintains better voice consistency. Supports both text and audio input/output.
Text to speech · Audio generation · Conversation
Released 2025-06-17Large model
OpenAI
Cost-efficient version of GPT Audio. Features an upgraded decoder for more natural sounding voices and better voice consistency at a fraction of the cost.
Text to speech · Audio generation · Conversation
Released 2025-06-17Medium model
OpenAI
codex-mini-latest is a fine-tuned version of o4-mini specifically for use in Codex CLI. For direct use in the API, we recommend starting with gpt-4.1.
Code generation · Conversation · Analysis
Released 2025-05-16Medium model
OpenAI
OpenAI's reasoning model for coding, math, science and visual perception. Thinks longer before answering and reaches for tools on its own: web search, code execution, image generation. Set new SOTA on Codeforces and MMMU.
Conversation · Reasoning · Code generation +2
Released 2025-04-16Extra large model
OpenAI
A smaller, cheaper OpenAI reasoning model built for speed. Strong for its size on math, coding and visual tasks, with much higher usage limits than o3, and it can use tools on its own.
Conversation · Reasoning · Code generation +1
Released 2025-04-16Medium model
OpenAI
The same model as o4-mini, defaulting to high reasoning effort. Compact, fast and cheap, keeping the multimodal and agentic behaviour, especially on math, coding and visual tasks.
Conversation · Reasoning · Code generation +1
Released 2025-04-16Medium model
OpenAI
OpenAI's GPT-4.1, built around a 1 million token context window and precise code diffs. Scores 54.6% on SWE-bench Verified and 87.4% on IFEval, and beats GPT-4o on coding, instruction compliance and long-document recall.
Conversation · Reasoning · Code generation +1
Released 2025-04-14Extra large model
OpenAI
The smallest and cheapest model in the GPT-4.1 line, with the same 1 million token context. Scores 80.1% on MMLU, 50.3% on GPQA and 9.8% on Aider polyglot, above GPT-4o mini. Aimed at classification and autocompletion.
Conversation · Reasoning · Code generation +1
Released 2025-04-14Small model
OpenAI
The mid tier of the GPT-4.1 line, roughly GPT-4o quality at lower latency and cost, with the same 1 million token context. Scores 84.1% on IFEval, 35.8% on MultiChallenge and 31.6% on Aider's polyglot diff benchmark, and handles vision.
Conversation · Analysis · Code generation
Released 2025-04-14Medium model
OpenAI
The March 2025 refresh of GPT-4o in ChatGPT. Follows instructions more closely, handles coding with fewer detours, and writes shorter answers with fewer nested markdown levels than the previous build.
Conversation · Reasoning · Analysis +1
Released 2025-03-27Extra large model
OpenAI
GPT-4o with audio input support. Detects nuances within audio recordings and adds depth to generated user experiences. Processes both text and audio prompts.
Text to speech · Audio generation · Conversation
Released 2025-03-20Large model
OpenAI
GPT-4.5 scaled up pre-training and post-training. Broader knowledge, better intent reading and more 'EQ' on conversation, writing, programming and practical problems, with fewer hallucinations. Scored 62.5% on SimpleQA with a 37.1% hallucination rate, ahead of GPT-4o.
Conversation · Reasoning · Code generation +2
Released 2025-02-27Extra large model
OpenAI
A smaller, cheaper o3 with faster response times. OpenAI says output quality holds.
Conversation · Reasoning · Code generation
Released 2024-12-15Small model
OpenAI
o1 achieves 86% accuracy on Mathematics Olympiad benchmarks (vs. GPT-4o's 13%), offers PhD-level STEM proficiency, and is built with extensive alignment training and safety evaluation.
Conversation · Web design · Code generation +1
Released 2024-12-05Medium model
OpenAI
OpenAI's small model after GPT-4 Omni: text and image in, text out. Many multiples cheaper than the frontier models of its day and over 60% cheaper than GPT-3.5 Turbo, at close to SOTA intelligence.
Conversation · Reasoning · Code generation +1
Released 2024-07-18Small model
OpenAI
GPT-4o processes text, images, and audio through a unified transformer architecture and offers real-time translation for 154 languages with 89.2% BLEU score on low-resource languages.
Conversation · Reasoning · Analysis +2
Released 2024-05-13Extra large model
OpenAI
OpenAI's high-definition text-to-speech with six built-in voices. Clear and natural, at the cost of latency.
Text to speech · Audio generation
Released 2024-04-09Large model
OpenAI
OpenAI's standard text-to-speech, tuned for low latency. Six voices, several output formats. Faster than TTS-1-HD and slightly rougher.
Text to speech · Audio generation
Released 2023-11-06Medium model
OpenAI
DALL-E 3 auto-improves user inputs via ChatGPT integration and blocks prohibited content with 99.9% precision using multimodal classifiers.
Image generation
Released 2023-09-01Large model
OpenAI
The original GPT-4: a large multimodal model that handled hard problems more accurately than GPT-3.5 on the strength of broader general knowledge and better reasoning. Training data stops at September 2021.
Conversation · Reasoning · Code generation +1
Released 2023-05-29Extra large model
OpenAI
OpenAI's fastest model of its generation. Understands and writes natural language and code, tuned for chat and completion.
Conversation · Code generation
Released 2022-03-15Medium model
OpenAI
A direct scale-up of GPT-1 with 1.5 billion parameters, trained on 8 million web pages. Known for its ability to generate coherent text, sometimes indistinguishable from humans, but could be repetitive.
Conversation · Summarization · Analysis
Released 2019-11-05Medium model
OpenAI
The first large-scale transformer-based language model released by OpenAI, trained on the BooksCorpus dataset. This version is accessed via the Hugging Face model hub (`openai-community/openai-gpt`).
Conversation · Summarization
Released 2018-06-11Small model
Google AI
Google's most capable Flash model at release, with gains over 3.7 Flash in software engineering, agent tasks and multi-step reasoning. It accepts text, images, audio, video and files across a 1M-token window, with up to 65,536 output tokens.
Conversation · Reasoning · Code generation +4
Released 2026-09-02Large model
Google AI
Google's Gemini 3.7 Flash, for fast agent workflows, coding and multi-step reasoning. Text, images, audio, video and files across a 1M-token context.
Conversation · Reasoning · Code generation +4
Released 2026-08-13Large model
Google AI
Google's Flash workhorse at the 3.6 mark, for coding, agent loops and web and app development. Built on 3.5 Flash, aiming at cleaner output with fewer unnecessary edits and less hedging, on roughly 17% fewer output tokens. Text, images, audio, video and PDFs across 1M tokens, up to 64K out.
Conversation · Reasoning · Code generation +4
Released 2026-07-21Large model
Google AI
Gemini 3.1 Flash Lite Image, the cheapest and fastest of Google's Nano Banana line. Text-to-image in roughly 4 seconds, about 2.7x faster than Gemini 3.1 Flash Image, with the same character consistency. One API for text-to-image, editing and composition. 1K resolution, 14 aspect ratios, SynthID watermark.
Image generation
Released 2026-06-30Small model
Google AI
Google's Flash tier at the 3.5 mark, close to Pro on coding and reasoning at Flash cost and speed. Tuned for coding and parallel agent loops, taking text, image, video, audio and PDF input. Thinking effort defaults to medium and runs from minimal to high.
Conversation · Reasoning · Code generation +3
Released 2026-05-19Large model
Google AI
The GA release of Gemini 3.1 Flash Lite, Google's cheapest multimodal tier. Text, image, video, audio and PDF in, for light agent work, simple extraction, and anything where latency and cost decide the design. Thinking runs minimal to high. Half the price of Gemini 3 Flash.
Conversation · Reasoning · Analysis +4
Released 2026-05-07Small model
Google AI
An instruction-tuned mixture-of-experts from Google DeepMind: 25.2B total parameters, 3.8B active per token, which is how it reaches near-31B quality on a fraction of the compute. Text, images and video up to 60 seconds at 1fps. 256K context, native function calling, thinking mode. Apache 2.0.
Conversation · Reasoning · Code generation +2
Released 2026-04-03Large model
Google AI
Google DeepMind's 30.7B dense multimodal model: text and image in, text out. 256K context, configurable thinking mode, native function calling, 140+ languages. Strongest on coding, reasoning and document understanding. Apache 2.0.
Conversation · Reasoning · Code generation +2
Released 2026-04-02Extra large model
Google AI
Google's high-efficiency tier at the 3.1 mark. Beats Gemini 2.5 Flash Lite on quality and gets close to 2.5 Flash, with the gains in audio and speech recognition, RAG snippet ranking, translation, extraction and code completion. 2.5x faster to first token than 2.5 Flash, at half the price of Gemini 3 Flash.
Conversation · Reasoning · Analysis +4
Released 2026-03-03Small model
Google AI
Gemini 3.1 Flash Image Preview, sold as Nano Banana 2. Google's image generation and editing model at Flash speed and Flash pricing, aimed at iterative edits rather than one-shot renders.
Image generation
Released 2026-02-26Large model
Google AI
The 3.1 update to Gemini Pro: better software engineering scores, more reliable agent behaviour, fewer tokens spent getting there. Reasons across text, image, video, audio and code over a 1M token context. Adds a medium thinking level, and improves on structured work like finance and spreadsheets.
Conversation · Reasoning · Code generation +3
Released 2026-02-19Extra large model
Google AI
Google DeepMind's Lyria 3, launched in the Gemini app. First Lyria with full vocal generation, automatic lyrics and multimodal input (text, image, video). Produces 30-second tracks with distinct verses, choruses and bridges, in 8 languages. No public API: it runs in the Gemini app and YouTube Dream Track.
Music generation · Audio generation
Released 2026-02-18Extra large model
Google AI
Gemini 3's Flash tier, for agent loops, multi-turn chat and coding assistance at much lower latency than Pro. 1M token context. Text, image, audio, video and PDF in, text out. Thinking level runs minimal to high, with structured output, tool use and automatic context caching.
Conversation · Reasoning · Code generation +3
Released 2025-12-17Large model
Google AI
Google DeepMind's production music generation model available via Vertex AI. Generates 32.8-second instrumental clips at 48kHz from text prompts. Supports negative prompts to exclude unwanted elements and deterministic generation via seed parameter. All output is SynthID-watermarked.
Music generation · Audio generation
Released 2025-12-10Large model
Google AI
Gemini 3 Pro Preview with high reasoning effort enabled. Exposes full chain-of-thought process for enhanced transparency in complex problem-solving across text, code, and multimodal tasks.
Conversation · Reasoning · Code generation +2
Released 2025-11-18Extra large model
Google AI
Google's Gemini 3 Pro Image model, sold as Nano Banana Pro. Generates and edits at up to 4K and routes prompts through Gemini's thinking mode for text rendering and multi-step instructions.
Image generation
Released 2025-11-12Large model
Google AI
The September 2025 checkpoint of Gemini 2.5 Flash, Google's workhorse tier for reasoning, coding, mathematics and science. Thinking is built in, and the max tokens for reasoning parameter caps how long it deliberates.
Conversation · Reasoning · Code generation +1
Released 2025-09-25Extra large model
Google AI
The September 2025 checkpoint of Gemini 2.5 Flash-Lite, the cheapest and lowest latency tier in the family. Thinking is off by default; turning it on through the Reasoning API parameter trades cost for intelligence.
Conversation · Reasoning · Code generation +1
Released 2025-09-25Large model
Google AI
Google's Nano Banana (Gemini 2.5 Flash Image) generates and edits images, tuned for speed. Tested on LMArena as 'nano-banana', where it beat Midjourney and Flux in blind tests.
Image generation
Released 2025-08-26Large model
Google AI
Gemini 2.5 Flash variant tuned for fast text-to-image generation and simple image edits.
Image generation
Released 2025-08-26Large model
Google AI
Google DeepMind's Gemma 3n at an effective 2B parameters inside a 6B architecture. The MatFormer design lets it nest submodels and recombine them through Mix-and-Match, which is how it runs on low-resource hardware. 32K context, multilingual.
Conversation · Reasoning · Translation
Released 2025-07-09Small model
Google AI
The lightweight tier of Gemini 2.5, tuned for low latency and cost rather than depth. Thinking is off by default to keep it fast; developers can switch it on through the Reasoning API parameter and pay for the extra intelligence.
Conversation · Reasoning · Analysis +1
Released 2025-06-17Small model
Google AI
Imagen 4 Fast trades some quality for speed and cost.
Image generation
Released 2025-06-12Large model
Google AI
Google's Gemini 2.5 Pro, a reasoning model that thinks through a request before answering. Aimed at coding, mathematics and scientific work, and it held first place on the LMArena leaderboard at release.
Conversation · Reasoning · Code generation +2
Released 2025-06-05Extra large model
Google AI
Imagen 4 Ultra trades speed and cost for image quality.
Image generation
Released 2025-05-21Large model
Google AI
Google's Imagen 4 text-to-image model, with strong detail and prompt adherence.
Image generation
Released 2025-05-21Large model
Google AI
The May 20 checkpoint of Gemini 2.5 Flash in its non-thinking form: it will not emit reasoning tokens, and is billed at the lower output rate. The paired :thinking variant is the one that reasons before answering.
Conversation · Reasoning · Code generation +1
Released 2025-05-20Medium model
Google AI
The May 20 checkpoint of Gemini 2.5 Flash with thinking enabled. It reasons before answering, is billed at the higher output rate, and takes a max tokens for reasoning parameter to cap how long it deliberates.
Conversation · Reasoning · Code generation +1
Released 2025-05-20Medium model
Google AI
Google's Gemma 3n at the E4B size, built to run on phones, laptops and tablets. Text, image and audio in, for transcription, translation and image analysis. Per-Layer Embedding caching and the MatFormer architecture activate only part of the model per request, which keeps memory use low. 32K context, 140+ languages.
Conversation · Analysis · Translation +1
Released 2025-05-20Small model
Google AI
Google's Gemini 2.5 Pro preview: multimodal understanding, multi-step code, math and STEM, and long-context analysis over large datasets, codebases and documents. Knowledge cutoff January 2025.
Conversation · Reasoning · Code generation +1
Released 2025-05-06Extra large model
Google AI
Google's Flash tier workhorse, with reasoning that can be switched on or off and a configurable thinking budget. The point is trading quality against cost and latency per request rather than per model.
Conversation · Reasoning · Code generation +1
Released 2025-04-17Medium model
Google AI
Gemini 2.5 Flash with thinking left on, so it reasons through a request before answering. The same workhorse model as the non-thinking variant, priced higher on output because thinking tokens are billed.
Conversation · Reasoning · Code generation +1
Released 2025-04-17Medium model
Google AI
Google's experimental Gemini 2.5 Pro: multimodal reasoning, a 1 million token context window, and video comprehension.
Conversation · Reasoning · Code generation +1
Released 2025-03-25Extra large model
Google AI
Gemma 3 adds multimodality: vision-language in, text out. 96,000 token context, 140+ languages, better math, reasoning and chat, with structured outputs and function calling. The 12B is the second largest, after the 27B.
Conversation · Reasoning · Code generation +1
Released 2025-03-13Large model
Google AI
Gemma 3 adds multimodality: vision-language in, text out. 131,072 token context, 140+ languages, better math, reasoning and chat, with structured outputs and function calling. The 27B is Google's largest Gemma 3, successor to Gemma 2.
Conversation · Reasoning · Code generation +1
Released 2025-03-12Extra large model
Google AI
A faster and cheaper Imagen 3 variant for when speed and cost matter more than maximum quality.
Image generation
Released 2025-02-06Large model
Google AI
Google's high-quality text-to-image model focused on lighting, detail, and strong visual composition.
Image generation
Released 2025-02-05Large model
Google AI
Gemini 2.0 Pro builds interactive 3D environments from text descriptions and offers hypothetical reasoning for scientific simulations.
Conversation · Reasoning · Analysis +2
Released 2025-01-01Extra large model
Google AI
An experimental Gemini 2.0 Flash variant that reasons through a problem before answering. Trades some latency for accuracy on multi-step questions while staying at Flash speed.
Conversation · Reasoning · Analysis +1
Released 2024-12-11Large model
Google AI
Gemini 1.5 Pro handles infinite context with 99% retrieval accuracy at 750k tokens via Mixture-of-Experts and generates chapter summaries for 2-hour videos with 92% accuracy.
Conversation · Reasoning · Analysis +1
Released 2024-02-15Large model
Google AI
Google's multimodal Gemini at release: natural language, multi-turn chat, code generation and image input.
Conversation · Reasoning · Code generation
Released 2023-12-13Large model
Google AI
Google's PaLM 2, with better multilingual, reasoning and coding work than PaLM. Tuned for chat.
Conversation · Reasoning · Code generation
Released 2023-07-20Large model
Anthropic
Anthropic's direct upgrade to Sonnet 5 for everyday coding and knowledge work. Anthropic says it writes output more than 30% faster than Sonnet 5 and finishes tasks with fewer tokens and tool calls, at the same price and a 1M-token context, with text, image and file input.
Conversation · Reasoning · Code generation +3
Released 2026-09-28Large model
Anthropic
Anthropic's successor to Opus 5 for reasoning, coding, and extended agent workflows. Accepts text, images, and documents in a one-million-token context, with adaptive thinking controlled by reasoning effort.
Conversation · Reasoning · Analysis +2
Released 2026-09-22Large model
Anthropic
Claude Fable 5.1 is a direct update to Fable 5, with its largest gains in agent coding, long-running workflows and knowledge work. Anthropic calls out long refactors, front-end generation, finance and analysis, with shorter plans and summaries than Fable 5.
Conversation · Reasoning · Code generation +5
Released 2026-09-01Extra large model
Anthropic
Anthropic's largest model, for demanding reasoning, coding and long-horizon agent work. Strongest on end-to-end software tasks, code review and bug finding, visual analysis of charts and documents, and coordinating parallel subagents. Instruction following holds across long tasks, even at lower effort.
Conversation · Reasoning · Code generation +3
Released 2026-07-24Extra large model
Anthropic
Anthropic's most capable Sonnet, with adaptive thinking across selectable effort levels from low to x-high, a 1M token context, and text, image and file input. It ships an updated tokenizer and real-time cyber safeguards that block certain high-risk dual-use requests.
Conversation · Reasoning · Code generation +3
Released 2026-06-30Large model
Anthropic
Anthropic's first public Mythos-class model, for autonomous knowledge work and coding. Text, image and file input, 1M token context. Built for long, ambiguous, multi-step tasks that used to need check-ins, self-correcting through verification loops. Falls back to Claude Opus 4.8 on high-risk queries.
Conversation · Reasoning · Code generation +4
Released 2026-06-09Extra large model
Anthropic
Anthropic's most capable generally available Opus. Text, image and file input, 1M token context. Built for autonomous agents and memory-driven work where coherence across a long session is the hard part: multi-step reasoning, complex coding, orchestration across large codebases, multi-stage debugging.
Conversation · Reasoning · Code generation +2
Released 2026-05-28Extra large model
Anthropic
Anthropic's Opus at the 4.7 mark, for long-running asynchronous agents. Stronger than Opus 4.6 on multi-step work: large codebases, multi-stage debugging, end-to-end orchestration. Beyond coding it drafts documents, builds presentations and analyses data, holding coherence across very long sessions.
Conversation · Reasoning · Code generation +2
Released 2026-04-15Extra large model
Anthropic
Anthropic's Sonnet 4.6, for coding, agents and professional work: iterative development, navigating large codebases, end-to-end project management with memory, document creation, and computer use for web QA and workflow automation.
Conversation · Reasoning · Code generation +3
Released 2026-02-17Large model
Anthropic
Anthropic's strongest model for coding and long-running professional work, built for agents that operate across a whole workflow rather than a single prompt. Handles large codebases, complex refactors and multi-step debugging, and produces near-final documents, plans and analyses in one pass.
Conversation · Reasoning · Code generation +2
Released 2026-02-04Extra large model
Anthropic
Anthropic's Opus tier at the 4.5 mark, aimed at long coding sessions and agent work. Text and image input, training data through August 2025, available on the Claude API, Bedrock and Vertex.
Conversation · Reasoning · Code generation +2
Released 2025-11-24Extra large model
Anthropic
Claude Haiku 4.5 matches Claude Sonnet 4's reasoning and coding strength while staying ultra-responsive for real-time and high-volume workloads. Extended thinking, controllable reasoning depth, and full tool-use support keep sub-agents fast without sacrificing quality.
Conversation · Reasoning · Code generation +2
Released 2025-10-15Large model
Anthropic
Anthropic's Sonnet 4.5, tuned for agents and coding workflows: speculative parallel execution, memory handling and tool orchestration.
Conversation · Reasoning · Code generation +2
Released 2025-09-29Large model
Anthropic
An update to Anthropic's Opus tier, sharpened on coding, reasoning and agent work. Scores 74.5% on SWE-bench Verified, with the clearest gains in multi-file refactors and debugging precision. Extended thinking runs to 64K tokens.
Conversation · Reasoning · Code generation +3
Released 2025-08-05Extra large model
Anthropic
The top tier of Anthropic's Claude 4 line, for coding, reasoning and agents. Built for long-running tasks, with extended thinking during tool use and better memory.
Conversation · Reasoning · Code generation +3
Released 2025-05-22Extra large model
Anthropic
Claude Sonnet 4 is a significant upgrade to Sonnet 3.7, delivering superior coding and reasoning while responding more precisely to instructions. It balances performance and efficiency for various use cases.
Conversation · Reasoning · Code generation +3
Released 2025-05-22Large model
Anthropic
Claude 3.7 Thinking Sonnet exposes the full chain-of-thought process during problem-solving, including error backtracking and alternative solution exploration. Scores 86.1% on GPQA Diamond benchmark for expert-level Q&A.
Conversation · Reasoning · Analysis +1
Released 2025-02-26Large model
Anthropic
Claude 3.7 Sonnet offers Extended Thinking Scaffolds that boost SWE-bench coding accuracy from 62.3% to 70.3%, with 81.2% accuracy in retail automation tasks, outperforming Claude Sonnet 3.6 (2022-10-22) by 13.6%.
Conversation · Reasoning · Analysis +1
Released 2025-02-25Large model
Anthropic
Claude 3.5 Sonnet offers a cost-efficient API ($3/million input tokens vs. $5 for GPT-4o) and uses embedded alignment techniques that reduce harmful outputs by 34% compared to Claude 2.1.
Conversation · Reasoning · Analysis +1
Released 2024-06-01Large model
Anthropic
A research demo of Claude 3 Sonnet, live for 24 hours on 23 May 2024, engineered by Anthropic to demonstrate feature steering. It was manipulated to obsess over the Golden Gate Bridge in every response.
Conversation · Reasoning
Released 2024-05-23Large model
Anthropic
The middle tier of Claude 3: a 200,000 token context window, priced for scaled deployment, and steady on conversation, reasoning, code generation and analysis.
Conversation · Reasoning · Code generation +1
Released 2024-03-05Large model
Anthropic
Claude 3 Haiku is Anthropic's fastest and most compact model, designed for near-instant responses in real-time applications with strong multilingual support.
Conversation · Analysis · Summarization
Released 2024-03-04Medium model
Anthropic
The top tier of Anthropic's Claude 3 line, for complex reasoning and open-ended problem solving.
Conversation · Reasoning · Code generation +1
Released 2024-03-04Extra large model
Anthropic
Anthropic's Claude 2 model, featuring a large 100K token context window and strong performance on various benchmarks. Known for helpful, honest, and harmless AI conversations.
Conversation · Reasoning · Analysis +1
Released 2023-06-11Large model
Amazon
Nova 2 Lite is a fast, cheap reasoning model for everyday workloads. Text, images and video in, text out. Aimed at document processing, extracting information from video, code generation, grounded answers and multi-step agent workflows.
Conversation · Reasoning · Code generation +3
Released 2025-12-02Medium model
Amazon
The top tier of Amazon's Nova line, for complex reasoning and as the teacher model when distilling your own.
Conversation · Reasoning · Code generation +1
Released 2025-10-31Large model
Baidu
Baidu's lightweight reasoning MoE model (21B total, 3B active per token) refined for deep multi-step reasoning across logic, math, science, and code. 128K context, Apache-2.0.
Conversation · Reasoning · Analysis +1
Released 2025-10-09Medium model
Baidu
Baidu's 300B-parameter text Mixture-of-Experts model, 47B active per token. 123K context, Apache-2.0. Part of the ERNIE 4.5 series.
Conversation · Reasoning · Analysis +1
Released 2025-06-30Extra large model
Black Forest Labs
The highest fidelity image model from Black Forest Labs for maximum detail and realism.
Image generation
Released 2025-12-16Large model
Black Forest Labs
FLUX 2 Flex adds guidance and step controls for more steerable generations.
Image generation
Released 2025-11-25Large model
Black Forest Labs
Black Forest Labs' FLUX 2 Dev model for high-quality generations with flexible controls.
Image generation
Released 2025-11-24Large model
Black Forest Labs
FLUX 2 Pro focuses on premium quality output with strong prompt adherence.
Image generation
Released 2025-11-14Large model
Black Forest Labs
Black Forest Labs' top FLUX tier for text-to-image, aimed at prompt adherence and image detail rather than speed. The reference model the rest of the FLUX line is measured against.
Image generation
Released 2024-11-12Large model
Black Forest Labs
The premium tier of FLUX Kontext for text-driven image editing, with better typography than the standard tier. You describe the change in words.
Image generation
Released 2024-11-12Large model
Black Forest Labs
A FLUX model for editing images by describing the change in words. Built for transformations that keep the rest of the frame consistent instead of regenerating the whole picture.
Image generation
Released 2024-11-12Large model
Black Forest Labs
A fast and cost-efficient FLUX model designed for high-throughput text-to-image generation.
Image generation
Released 2024-07-30Large model
Bria
A text-to-image model trained entirely on licensed data, so it is commercial-ready. 4B parameters, with aesthetics and text rendering evaluated on par with much larger models.
Image generation
Released 2025-12-04Large model
Bria
SOTA open source model trained on licensed data, transforming intent into structured control for precise, high-quality AI image generation in enterprise and agentic workflows.
Image generation
Released 2025-12-04Large model
ByteDance
Seed 2.1 Turbo is a multimodal model from ByteDance Seed for coding and long-horizon agent workflows. It is suited for end-to-end software delivery, multi-step task execution, and understanding visual and video content, with capabilities for planning, debugging, and self-correction.
Conversation · Reasoning · Code generation +3
Released 2026-08-12Large model
ByteDance
ByteDance Seed's coding-agent model. Suited to frontend work, multilingual programming, and coding agents such as Claude Code, Kilo and OpenCode.
Code generation · Conversation · Reasoning +2
Released 2026-08-12Large model
ByteDance
ByteDance's mid tier Seed model for high-frequency enterprise work, ahead of Seed 1.8 overall. Reasoning effort is configurable, and it handles multimodal input and structured output. The target jobs are unstructured information processing, text creation, search, recommendation and data analysis.
Conversation · Reasoning · Code generation +2
Released 2026-03-10Medium model
ByteDance
ByteDance's Seedream 5.0 Lite is a unified multimodal image generation model with chain-of-thought reasoning and real-time web search, producing native 2K/4K images with accurate multi-language text rendering.
Image generation
Released 2026-02-13Medium model
ByteDance
ByteDance's Seedream 4.5 text-to-image model designed for strong aesthetics and composition.
Image generation
Released 2025-12-03Large model
ByteDance
Unified text-to-image generation and precise single-sentence editing at up to 4K resolution by ByteDance.
Image generation
Released 2025-11-15Large model
Cohere
Cohere's model for enterprise agent workflows, with text and image inputs and a 192,000-token context. Supports tool calling, strict tool schemas, and structured output.
Conversation · Reasoning · Analysis +2
Released 2026-09-22Large model
Cohere
Cohere's first agentic coding model and the debut of its North family. A sparse mixture of experts, 30B total and 3B active, for code generation, agentic software engineering and terminal work, trained to generalise across scaffolds like OpenCode and SWE-Agent. 256K context, 64K output. Apache 2.0.
Conversation · Reasoning · Code generation +1
Released 2026-06-17Small model
DeepSeek
DeepSeek V4.1 Flash supports text and image inputs with text output and a 1M context window. OpenRouter token pricing varies by provider and time of use.
Conversation · Reasoning · Tool use
Released 2026-09-10Large model
DeepSeek
Experimental vision-enabled version of DeepSeek V4 Flash 0731, adding image understanding to the base model. Supports text and image inputs with a 1M context window on OpenRouter.
Conversation · Reasoning · Tool use
Released 2026-08-21Large model
DeepSeek
DeepSeek's large mixture-of-experts model and the general-availability release of V4 Pro. Keeps the 1M-token context window and hybrid attention of the V4 line, aimed at reasoning, coding and long-horizon agent work.
Conversation · Reasoning · Code generation +2
Released 2026-08-12Extra large model
DeepSeek
DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision of DeepSeek V4 Flash is suited for coding, reasoning, and agent workflows.
Conversation · Reasoning · Code generation +2
Released 2026-07-31Large model
DeepSeek
DeepSeek V4 Flash, a mixture-of-experts model with 284B total and 13B activated parameters over a 1M token context. Hybrid attention handles the long context, reasoning modes are configurable, and the whole thing is tuned for throughput: coding assistants, chat and agent loops.
Conversation · Reasoning · Code generation +2
Released 2026-04-24Large model
DeepSeek
DeepSeek V4 Pro, a mixture-of-experts model with 1.6T total and 49B activated parameters over a 1M token context. The same hybrid attention architecture as V4 Flash at much larger scale, with multiple reasoning modes and strong knowledge, mathematics and software engineering scores.
Conversation · Reasoning · Code generation +2
Released 2026-04-24Extra large model
DeepSeek
The high-compute variant of DeepSeek V3.2, with reinforcement learning post-training scaled past the base model. It keeps DeepSeek Sparse Attention for long context, and DeepSeek reports it ahead of GPT-5 on hard reasoning workloads and comparable to Gemini 3.0 Pro, with coding and tool use intact.
Conversation · Reasoning · Code generation +2
Released 2025-12-01Large model
DeepSeek
DeepSeek V3.2 introduces DeepSeek Sparse Attention, cutting training and inference cost without giving up long-context quality. Reinforcement learning post-training puts its reported reasoning in the GPT-5 class, with gold-medal results on the 2025 IMO and IOI, plus better tool-use compliance.
Conversation · Reasoning · Code generation +2
Released 2025-12-01Large model
DeepSeek
DeepSeek-V3.2-Exp introduces DeepSeek Sparse Attention (DSA) for efficient long-context. Reasoning toggle supported via boolean flag.
Conversation · Reasoning · Code generation +1
Released 2025-09-29Large model
DeepSeek
DeepSeek V3.1 runs one set of weights in two modes: deepseek-chat answers directly, deepseek-reasoner thinks first. Open weights on Hugging Face, 160K context, tuned for agentic tool calling.
Conversation · Reasoning · Code generation +4
Released 2025-08-21Extra large model
DeepSeek
DeepSeek R1 0528 is the May 28th update to the original DeepSeek R1. Performance on par with OpenAI o1, but open-sourced and with fully open reasoning tokens. It's 671B parameters in size, with 37B active in an inference pass. Fully open-source model.
Conversation · Reasoning · Code generation +1
Released 2025-05-28Extra large model
DeepSeek
A 671B parameter model, speculated to be geared towards logic and mathematics. Likely an upgrade from DeepSeek-Prover-V1.5. Released on Hugging Face without an announcement or description.
Reasoning · Analysis · Conversation +1
Released 2025-04-30Extra large model
DeepSeek
DeepSeek R1 is a reasoning model developed entirely via reinforcement learning, offering cost efficiency at $0.14/million tokens vs. OpenAI o1's $15, with strong code generation and analysis capabilities.
Conversation · Reasoning · Code generation +1
Released 2025-02-01Large model
DeepSeek
DeepSeek V3 (March 2024) shows significant improvements in reasoning capabilities with enhanced MMLU-Pro (81.2%), GPQA (68.4%), AIME (59.4%), and LiveCodeBench (49.2%) scores. Features improved front-end web development, Chinese writing proficiency, and function calling accuracy.
Conversation · Reasoning · Web design +2
Released 2024-03-24Extra large model
ElevenLabs
Expressive TTS with emotion, laughter and breathing. ElevenLabs' top voice-quality tier.
Text to speech · Audio generation
Released 2025-10-15Large model
ElevenLabs
TTS at roughly 75ms latency, for real-time apps and voice agents.
Text to speech · Audio generation
Released 2025-06-20Medium model
ElevenLabs
High-quality multilingual TTS supporting 29 languages with voice cloning capabilities.
Text to speech · Audio generation · Translation
Released 2024-03-15Large model
Ideogram
Turbo is the fastest and cheapest Ideogram v3. v3 creates images with stunning realism, creative designs, and consistent styles.
Image generation
Released 2024-11-12Large model
Ideogram
The slowest and most expensive Ideogram v3 variant, and the most detailed. Pick it when quality outranks speed and cost.
Image generation
Released 2024-11-12Large model
Ideogram
The middle Ideogram v3 variant: text rendering and prompt comprehension at a moderate generation time.
Image generation
Released 2024-11-12Large model
Inception
Inception Mercury 2.5 is a diffusion-based language model for text and code generation, with a 260K context window on OpenRouter.
Conversation · Reasoning · Code generation +1
Released 2026-09-08Large model
Inception
Inception's diffusion language model generates and refines tokens in parallel instead of one at a time. It reaches a reported 1,107 tokens per second on standard GPUs, with a 260K-token context window and a claimed 10-point intelligence gain over Mercury 2.
Conversation · Reasoning · Code generation +3
Released 2026-08-31Medium model
Inception
Inception's Mercury 2, the first reasoning diffusion language model. Instead of emitting tokens one at a time it produces and refines many in parallel, clearing 1000 tokens per second on standard GPUs, over 5x faster than Claude 4.5 Haiku and GPT-5 Mini. 128K context, native tool use, schema-aligned JSON.
Conversation · Reasoning · Code generation +2
Released 2026-03-04Medium model
Inception
The first diffusion large language model. Discrete diffusion runs 5 to 10x faster than speed-tuned models like GPT-4.1 Nano and Claude 3.5 Haiku while matching them, which is what makes voice agents and search viable.
Conversation · Reasoning · Code generation +1
Released 2025-06-26Medium model
inclusionAI
Vision-language variant of inclusionAI Ling 3.0 Flash, adding image and video understanding to its mixture-of-experts language model. Available through a free OpenRouter endpoint with a 262K context window.
Conversation · Reasoning · Tool use
Released 2026-09-10Large model
inclusionAI
Health and medicine-focused variant of inclusionAI Ling 3.0 Flash, with 124B total and 5.1B active parameters. The free OpenRouter endpoint supports a 262K context window.
Conversation · Reasoning · Tool use
Released 2026-09-04Large model
inclusionAI
Finance-focused variant of inclusionAI Ling 3.0 Flash, with 124B total and 5.1B active parameters, designed for investment research and financial analysis. Available through a free 262K-context OpenRouter endpoint.
Conversation · Reasoning · Tool use
Released 2026-08-27Large model
inclusionAI
inclusionAI mixture-of-experts model with 124B total and about 5.1B active parameters, designed for token-efficient agentic inference at production scale.
Conversation · Reasoning · Tool use
Released 2026-07-23Extra large model
inclusionAI
inclusionAI's Ring-2.6-1T, a trillion-parameter thinking model with 63B active, for agent workflows that must be capable and cheap. Tuned for coding agents, tool use and long-horizon execution. inclusionAI reports leading results on PinchBench, ClawEval, TAU2-Bench and GAIA2-search.
Conversation · Reasoning · Code generation +2
Released 2026-05-08Extra large model
inclusionAI
inclusionAI's trillion-parameter Ling, using a fast-thinking approach that cuts cost to roughly a quarter of comparable models. inclusionAI reports leading results on AIME26 and SWE-bench Verified, aimed at coding, complex reasoning and large-scale agent workflows.
Conversation · Reasoning · Code generation +2
Released 2026-04-23Extra large model
inclusionAI
inclusionAI's Ling-2.6-flash, 104B total parameters with 7.4B active, built for agents that need fast answers and few tokens. inclusionAI reports it level with comparable models at similar scale while spending noticeably less on coding, document processing and light agent work.
Conversation · Reasoning · Code generation +1
Released 2026-04-21Large model
Luma
Luma's image model for creative professional work. Photorealistic output with careful lighting, material rendering and composition.
Image generation
Released 2025-03-01Large model
Luma
Accelerated variant of Luma Photon prioritizing speed while maintaining quality. Ideal for rapid iteration and production workflows where faster generation times are critical.
Image generation
Released 2025-03-01Large model
Meta AI
Meta's lower-cost Muse Spark 1.3 tier for experiments, learning, coding and early agent workflows. It follows information across long tasks and asks for confirmation when inputs conflict. Access is restricted to adults; prompts and outputs may be used to improve Meta's products.
Conversation · Reasoning · Code generation +5
Released 2026-09-02Large model
Meta AI
Meta's multimodal reasoning model for long-running agent and coding workflows. It tracks state across extended tasks, works through conflicting inputs and asks for confirmation when needed, with a 1M-token context window and concise execution. Access is restricted to adults.
Conversation · Reasoning · Code generation +5
Released 2026-09-02Large model
Meta AI
Meta Superintelligence Labs dense, open-weight multimodal model distilled from Muse Spark and designed for autonomous agents on consumer hardware. Supports text and image inputs with a 128K context window on OpenRouter.
Conversation · Reasoning · Tool use
Released 2026-08-09Medium model
Meta AI
Meta's multimodal reasoning model for agent work. Text, images, video, audio and PDFs in, text out, over a 1M token context. Orchestrates multi-agent workflows as planner or subagent, and generalises zero-shot to new tools, MCP servers and custom skills. Parallel function calling, search with citations.
Conversation · Reasoning · Code generation +4
Released 2026-07-09Large model
Meta AI
Meta's multimodal MoE: 17B active parameters across 128 experts, 400B total. Beat GPT-4o and Gemini 2.0 Flash on several benchmarks, with an ELO of 1417 on LMArena. Built for image understanding and creative writing.
Conversation · Reasoning · Code generation +2
Released 2025-04-05Extra large model
Meta AI
Meta's compact multimodal MoE: 17B active parameters across 16 experts, 109B total. Fits on one H100 at Int4 quantization, carries a 10M token context window, and beat Gemma 3, Gemini 2.0 Flash-Lite and Mistral 3.1 on several benchmarks.
Conversation · Reasoning · Code generation +1
Released 2025-04-05Medium model
Meta AI
Llama 3.1 70B: a much larger context window than 3.0, and better math and general knowledge.
Conversation · Reasoning · Code generation
Released 2024-07-23Large model
Meta AI
Meta's largest open-weight Llama 3.1, which beat several proprietary models on benchmarks at release.
Conversation · Reasoning · Code generation +1
Released 2024-07-23Extra large model
Meta AI
Llama 3 70B is a large language model from Meta with strong performance and efficiency for real-time interactions.
Conversation · Reasoning · Code generation
Released 2024-04-18Large model
Meta AI
Meta's text-to-music generation model producing high-quality, diverse music from text descriptions. Trained on licensed music.
Music generation · Audio generation
Released 2023-06-12Large model
Midjourney
Midjourney v6 produces realistic images.
Image generation
Released 2024-12-01Large model
Midjourney
Midjourney v6.1 introduced a native web interface alongside Discord, with improved detail rendering, better text handling, and enhanced image coherence.
Image generation
Released 2024-07-31Large model
Midjourney
Midjourney v5 produces realistic images.
Image generation
Released 2023-11-01Large model
Midjourney
Midjourney v4 was the jump in photorealism, coherence and prompt understanding. First version trained on Google TPUs.
Image generation
Released 2022-11-05Large model
Midjourney
Midjourney v3 read prompt nuance and artistic style better than v2.
Image generation
Released 2022-07-25Large model
Midjourney
Midjourney v2 improved on the original model with better coherence, detail, and more consistent style application.
Image generation
Released 2022-04-12Medium model
Midjourney
The first public release of Midjourney, introducing AI image generation to a wider audience through its Discord-based interface.
Image generation
Released 2022-02-01Medium model
MiniMax
MiniMax M3 takes text, image and video, returns text, over a 1M token context, for long-horizon agent work, coding and tool use. MiniMax Sparse Attention swaps full attention for KV-block selection, cutting per-token compute at long context to roughly a twentieth of the last generation.
Conversation · Reasoning · Code generation +4
Released 2026-06-01Large model
MiniMax
MiniMax M2.7 runs multi-agent collaboration internally, planning, executing and refining across changing environments. Handles live debugging, root cause analysis, financial modelling and document generation across Word, Excel and PowerPoint. Scores 56.2% on SWE-Pro, 57.0% on Terminal Bench 2, 1495 ELO on GDPval-AA.
Conversation · Reasoning · Code generation +4
Released 2026-03-18Large model
MiniMax
MiniMax M2.5 extends M2.1's coding into office work: it generates and operates Word, Excel and PowerPoint files, switches software environments, and works alongside other agents. Scores 80.2% on SWE-Bench Verified, 51.3% on Multi-SWE-Bench and 76.3% on BrowseComp, on fewer tokens than earlier generations.
Conversation · Reasoning · Code generation +4
Released 2026-02-12Large model
MiniMax
A dialogue-first model for roleplay, character chat and long multi-turn conversation. Holds tone and personality, supports rich message roles, and can learn pacing from example dialogue.
Conversation · Reasoning
Released 2026-01-27Large model
MiniMax
MiniMax's 230B mixture-of-experts model with open weights under MIT, tuned for coding and agent work. Available from MiniMax's own platform, Hugging Face and OpenRouter.
Conversation · Reasoning · Code generation +1
Released 2025-12-23Large model
MiniMax
MiniMax M2: 10B activated parameters for coding agents, compile-run-fix loops and long-horizon reasoning. Good SWE-Bench and Terminal-Bench results, and it holds reasoning continuity across multi-step tasks.
Conversation · Reasoning · Code generation +4
Released 2025-10-23Medium model
MiniMax
MiniMax M1, an open-weight reasoning model with 456B total parameters and 45.9B active per token. It pairs a mixture-of-experts backbone with a lightning attention mechanism to hold context up to 1 million tokens without the usual FLOP cost, and is tuned for multi-step reasoning.
Conversation · Reasoning · Code generation +3
Released 2025-06-17Extra large model
MiniMax
MiniMax's first image generation model with character reference support for consistent character generation across multiple images. Produces high-quality, detailed images with strong compositional abilities.
Image generation
Released 2025-06-01Large model
MiniMax
High-definition text-to-speech with studio-quality output and fine-grained emotion control.
Text to speech · Audio generation
Released 2025-03-10Large model
Mistral AI
Mistral Small 4 unifies the capabilities of Magistral (reasoning), Pixtral (multimodal), and Devstral (agentic coding) into a single efficient model. It handles complex analysis, software development, and visual tasks within the same workflow at $0.15/$0.60 per million tokens.
Conversation · Analysis · Code generation +2
Released 2026-03-16Small model
Mistral AI
Mistral Small Creative is an experimental small model designed for creative writing, narrative generation, roleplay and character-driven dialogue, general-purpose instruction following, and conversational agents.
Conversation · Analysis
Released 2025-12-16Small model
Mistral AI
The free OpenRouter tier of Devstral 2, Mistral's 123B coding model. Open weights under a modified MIT licence, also sold first-party and through NVIDIA NIM.
Conversation · Reasoning · Code generation +1
Released 2025-12-09Extra large model
Mistral AI
The December 2025 build of Mistral Large 3, a 675B instruct model with open weights on Hugging Face. Sold first-party through Mistral AI Studio and resold on Bedrock, Azure and watsonx.
Conversation · Reasoning · Code generation +1
Released 2025-12-01Extra large model
Mistral AI
An update to Mistral Medium 3, pitched at large-model quality for roughly an eighth of the cost. Strongest on coding and STEM reasoning, competitive with Claude Sonnet 3.7 and Llama 4 Maverick, and deployable hybrid, on-prem or inside a VPC.
Conversation · Reasoning · Code generation +1
Released 2025-08-13Large model
Mistral AI
A code generation and agentic reasoning model built jointly by Mistral AI and All Hands AI, a step up from Devstral Small. Scores 61.6% on SWE-Bench Verified, ahead of Gemini 2.5 Pro and GPT-4.1 on code tasks at a fraction of the price.
Conversation · Reasoning · Code generation +1
Released 2025-07-11Medium model
Mistral AI
Devstral Small 1.1, a 24B open-weight model for software engineering agents, from Mistral AI with All Hands AI. Finetuned from Mistral Small 3.1, Apache 2.0, 128k context, with Mistral-style function calling and XML output.
Conversation · Reasoning · Code generation +1
Released 2025-07-11Small model
Mistral AI
Mistral's mid tier model, pitched at large-model quality for roughly an eighth of the cost. Strongest on coding and STEM reasoning, and deployable hybrid, on-prem or inside a VPC rather than API only.
Conversation · Reasoning · Code generation +1
Released 2025-05-07Large model
Mistral AI
Mistral Large 2 features a 128K context window with enhanced code generation, mathematics, reasoning, and multilingual support.
Conversation · Reasoning · Code generation
Released 2024-07-24Large model
Mistral AI
Mistral Neom 3 is a 12B parameter model with a 128k token context length built by Mistral in collaboration with NVIDIA.
Conversation · Code generation
Released 2024-07-19Large model
Mistral AI
Mistral Large: strong multilingual work and reasoning, with a 32K token context window.
Conversation · Reasoning · Code generation
Released 2024-02-26Large model
Moonshot AI
Moonshot AI's open-weight multimodal reasoning model at very large scale. Built for complex coding, knowledge work and long agent runs, strongest at navigating large repositories and iterating against images, logs, tests and runtime feedback. Uses KDA and Attention Residuals for efficiency.
Conversation · Reasoning · Code generation +3
Released 2026-07-16Extra large model
Moonshot AI
The coding model in Moonshot AI's K2.7 line, built to finish programming tasks end to end over long contexts. A native multimodal mixture of experts taking text, image and video, always in thinking mode, with the full reasoning kept across turns. 256K context, 32B active parameters out of roughly 1T.
Conversation · Reasoning · Code generation +3
Released 2026-06-13Extra large model
Moonshot AI
Moonshot AI's K2.6, aimed at long-horizon coding, UI generation from prompts or images, and multi-agent orchestration. It works end to end across Python, Rust and Go, and its agent swarm scales to hundreds of parallel sub-agents, returning documents, websites and spreadsheets in one run without supervision.
Conversation · Reasoning · Code generation +3
Released 2026-04-20Large model
Moonshot AI
Moonshot AI's natively multimodal K2.5, continuing pretraining on Kimi K2 over roughly 15T mixed visual and text tokens. Its two selling points are visual coding and a self-directed agent swarm, alongside general reasoning and tool calling.
Conversation · Reasoning · Code generation +1
Released 2026-01-27Large model
Moonshot AI
Kimi Linear is a hybrid linear attention architecture that outperforms traditional full attention methods. Features Kimi Delta Attention (KDA) for efficient memory usage, reducing KV caches by up to 75% and boosting throughput by up to 6x for contexts as long as 1M tokens.
Conversation · Reasoning · Code generation +1
Released 2025-11-10Extra large model
Moonshot AI
The reasoning model in Moonshot AI's K2 line, on the same trillion-parameter mixture-of-experts backbone with 32 billion active per pass and 256K context. It interleaves thinking with tool calls, and is built to hold a research or coding task together across hundreds of sequential actions without drifting.
Conversation · Reasoning · Code generation +2
Released 2025-11-06Extra large model
Moonshot AI
The September refresh of Kimi K2, still 1 trillion total parameters with 32 billion active per pass, now holding 256K context instead of 128K. The update targets agentic coding accuracy across different scaffolds and front-end output worth looking at, for web and 3D work.
Conversation · Reasoning · Code generation +2
Released 2025-09-04Extra large model
Moonshot AI
Moonshot AI's Kimi K2, a mixture-of-experts model with 1 trillion total parameters and 32 billion active per forward pass. Tuned for agentic work: tool use, reasoning and code synthesis, with 128K context. Trained with the MuonClip optimizer, which is what kept a MoE run at this scale stable.
Conversation · Reasoning · Code generation +1
Released 2025-07-11Extra large model
Nex AGI
Open-weight Nex-N2.5 Mini model for agentic coding with visual feedback, repository exploration, and multi-file implementation. Accepts text and images with a 262K context window on its free OpenRouter endpoint.
Conversation · Reasoning · Code generation +1
Released 2026-09-08Large model
Nex AGI
Open-weight Nex-N2.5 Pro model for agentic coding with visual feedback, repository exploration, and multi-file implementation. Accepts text and images with a 262K context window on its free OpenRouter endpoint.
Conversation · Reasoning · Code generation +1
Released 2026-09-08Large model
NVIDIA
NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that benefit from domain-specific customization. A free-tier endpoint is also available on OpenRouter.
Conversation · Reasoning · Code generation +2
Released 2026-08-11Medium model
NVIDIA
NVIDIA's open reasoning and orchestration model, 55B active parameters out of 550B on a hybrid Transformer-Mamba mixture of experts. Text in, text out, with a context window up to 1M tokens. Built for long-running agent work: orchestration, coding agents, deep research and multi-step reasoning.
Conversation · Reasoning · Code generation +4
Released 2026-06-04Extra large model
NVIDIA
A 4B multimodal guardrail model from NVIDIA, fine-tuned from Gemma-3-4B. It moderates both the input to an LLM and the response, taking text and images and returning a safe or unsafe classification for each, category labels, and an optional reasoning trace. 12 languages, 128K context.
Analysis · Data extraction
Released 2026-06-04Small model
NVIDIA
NVIDIA open multimodal reasoning model with 30B total and 3B active parameters, designed for perception and context processing in enterprise agents. Accepts text, image, audio, and video inputs with a 256K context window.
Conversation · Reasoning · Tool use
Released 2026-04-28Large model
NVIDIA
NVIDIA's Nemotron 3 Super: a 120B open hybrid mixture-of-experts activating 12B parameters per token. The Mamba-Transformer backbone with multi-token prediction generates over 50% more tokens than comparable open models, and Latent MoE calls 4 experts for the cost of one. 1M context, NVIDIA Open License.
Conversation · Reasoning · Code generation +3
Released 2026-03-11Large model
NVIDIA
NVIDIA trained Nemotron Nano 9B v2 from scratch as one model for both reasoning and non-reasoning work. It can expose an internal reasoning trace before the answer, or be told by system prompt to give the answer alone.
Conversation · Reasoning · Code generation +1
Released 2025-09-05Medium model
OpenRouter
A cloaked reasoning model on OpenRouter, out for community feedback. Built for coding, sustained agent work and production loads: long-horizon software engineering, complex reasoning, and multimodal workflows mixing text with images and video. Reasoning always on. 1M token context, 131K output.
Conversation · Reasoning · Code generation +4
Released 2026-08-24Large model
OpenRouter
OpenRouter Fusion, Quality preset, June 2026 snapshot. One prompt goes in parallel to Claude Opus 4.8, GPT-5.5 and Gemini 3 Pro with web search on, then Claude Opus 4.8 judges their answers into one. A request costs every panel member plus the judge. The panel drifts over time.
Conversation · Reasoning · Analysis +1
Released 2026-06-13Extra large model
OpenRouter
OpenRouter Fusion, Budget preset, June 2026 snapshot. The same deliberation on cheaper members: Gemini 3 Flash, DeepSeek V3.2 and Kimi K2.7 Code answer in parallel with web search on, then Gemini 3 Flash judges. Still costs every member plus the judge, far less than Quality.
Conversation · Reasoning · Analysis +1
Released 2026-06-13Extra large model
OpenRouter
A stealth foundation model for agent workloads, with native tool use and long context. Strong on code generation, automated workflows and complex instruction execution, and compatible with Claude Code, OpenClaw and similar tools. Prompts and completions may be logged by the provider.
Conversation · Reasoning · Code generation +3
Released 2026-04-28Large model
OpenRouter
Elephant Alpha is a 100B-parameter text model focused on intelligence efficiency, delivering strong reasoning performance while minimizing token usage. It supports a 256K context window with up to 32K output tokens, function calling, structured output, and prompt caching.
Conversation · Reasoning · Code generation +3
Released 2026-04-13Large model
OpenRouter
A stealth model on OpenRouter with roughly 1 trillion parameters and 1M tokens of context, built for agent work: long-horizon planning, multi-step execution, and the instruction-following precision frameworks like OpenClaw need. Prompts and completions are logged by the provider and may be used for training.
Conversation · Reasoning · Code generation +3
Released 2026-03-11Extra large model
OpenRouter
A stealth omni-modal model on OpenRouter that takes visual and audio input natively, reasons across both, and acts on the result through multi-step tasks. Prompts and completions are logged by the provider and may be used for training.
Conversation · Reasoning · Code generation +3
Released 2026-03-11Extra large model
OpenRouter
A cloaked reasoning model from OpenRouter, out for community feedback. Built for speed: coding assistants, real-time conversation and agent workflows. Default reasoning effort is medium; high is recommended for agentic coding.
Conversation · Reasoning · Code generation +2
Released 2026-02-09Large model
OpenRouter
A stealth model released through OpenRouter to collect community feedback, strong on coding, agent workflows, reasoning and roleplay. 200K context, with tool calling, structured outputs and step-by-step reasoning.
Conversation · Code generation · Reasoning +3
Released 2026-02-06Large model
OpenRouter
Z.ai's 358B open-weight model under MIT, aimed at coding and agent loops. Runs on the first-party Z.ai API, on OpenRouter, and on Cerebras for high-speed inference.
Conversation · Reasoning · Code generation +1
Released 2025-12-22Large model
OpenRouter
TNG's November 2025 rebuild of its R1T Chimera merge, a DeepSeek R1 derivative. It runs only on OpenRouter's free tier, with no separate weights repository published.
Conversation · Reasoning
Released 2025-11-27Large model
OpenRouter
Prime Intellect's 106B mixture-of-experts model, 12B active per token, post-trained from GLM-4.5-Air-Base with supervised fine-tuning and reinforcement learning. Open weights under MIT.
Conversation · Reasoning · Code generation +1
Released 2025-11-27Large model
OpenRouter
A cloaked OpenRouter endpoint from late November 2025, later revealed as an early build of Mistral Large 3. Free while it ran, prompts logged by the provider, now retired.
Conversation · Reasoning · Analysis
Released 2025-11-24Large model
OpenRouter
A cloaked OpenRouter endpoint from November 2025, later revealed as an early snapshot of xAI's Grok 4.1 Fast with reasoning switched off. Free while it ran, prompts logged by the provider.
Conversation · Reasoning · Code generation +2
Released 2025-11-15Extra large model
OpenRouter
The thinking half of OpenRouter's Sherlock alpha pair, later revealed as an early snapshot of xAI's Grok 4.1 Fast with reasoning switched on. Free while it ran, prompts logged by the provider.
Conversation · Reasoning · Code generation +2
Released 2025-11-15Extra large model
OpenRouter
A general-purpose model, strongest on coding, tool calling and instruction following.
Conversation · Reasoning · Code generation +2
Released 2025-11-06Large model
OpenRouter
Smaller reasoning model trained for image and visual understanding. Cloaked trial; prompts and outputs logged for provider feedback.
Conversation · Reasoning · Analysis
Released 2025-10-21Medium model
OpenRouter
A cloaked model OpenRouter ran for community feedback: fast, general purpose, with a 2 million token context window. Takes image input and calls tools in parallel.
Conversation · Reasoning · Code generation +1
Released 2025-09-05Extra large model
OpenRouter
A cloaked model OpenRouter ran in September 2025 with a two million token context. It was later revealed as the reasoning variant of xAI's Grok 4 Fast, and the alpha endpoint has since been retired.
Conversation · Reasoning · Code generation +3
Released 2025-09-05Extra large model
OpenRouter
This is a cloaked model provided to the community to gather feedback. This is an improved version of Horizon Alpha. Note: It's free to use during this testing period, and prompts and completions are logged by the model creator for feedback and training.
Conversation · Reasoning · Code generation +1
Released 2025-08-01Large model
OpenRouter
This is a cloaked model provided to the community to gather feedback. Note: It's free to use during this testing period, and prompts and completions are logged by the model creator for feedback and training.
Conversation · Reasoning · Code generation +1
Released 2025-07-30Large model
OpenRouter
A cloaked model OpenRouter offered free in July 2025 to gather feedback, with a one million token context. The lab behind it was never publicly confirmed, prompts were logged, and the endpoint is now retired.
Conversation · Reasoning · Code generation +1
Released 2025-07-01Large model
OpenRouter
A stealth general-purpose model for long-context work, including code generation. Shipped on community feedback.
Conversation · Web design · Code generation +1
Released 2025-04-10Large model
OpenRouter
A cloaked model released for community feedback. General-purpose, handles long context and code generation. Every prompt and completion is logged by the provider and by OpenRouter.
Conversation · Web design · Code generation +1
Released 2025-04-02Large model
OpenRouter
Open-source zero-shot text-to-speech with natural prosody. Generates highly natural speech from just a reference audio clip.
Text to speech · Audio generation
Released 2024-10-08Medium model
OpenRouter
Open-source audio model: speech, music and sound effects, including laughter, crying and other non-verbal sounds.
Text to speech · Audio generation · Music generation
Released 2023-04-20Medium model
Poolside
Poolside open-weight coding model with 118B total and 8B active parameters, designed for software engineering and agent workflows. This free OpenRouter endpoint exposes a 262K context window.
Conversation · Reasoning · Code generation +1
Released 2026-07-21Large model
Poolside
Poolside open-weight coding model with 33B total and 3B active parameters, designed for software engineering and agent workflows. This free OpenRouter endpoint exposes a 262K context window.
Conversation · Reasoning · Code generation +1
Released 2026-07-02Large model
PrunaAI
A sub 1-second text-to-image model built for production use cases by PrunaAI.
Image generation
Released 2026-01-13Large model
PrunaAI
A turbo text-to-image model optimized by PrunaAI for very fast inference at low guidance.
Image generation
Released 2025-11-26Large model
PrunaAI
PrunaAI-optimized HiDream L1 Fast for cheap, fast text-to-image generation with configurable speed modes.
Image generation
Released 2025-04-10Large model
Qwen
A higher-throughput version of Qwen3.8 Max, offered as a separate endpoint at a higher price. It accepts text, images, and video with a one-million-token context and configurable reasoning effort.
Conversation · Reasoning · Analysis +2
Released 2026-09-23Large model
Qwen
Alibaba's Qwen3.8 reasoning model with native audio and video understanding. It accepts text, images, audio, and video, returns text, and supports a one-million-token context for analysis and agent workflows.
Conversation · Reasoning · Analysis +2
Released 2026-09-21Large model
Qwen
September 2 snapshot of Qwen3.8 Max, supporting text, image, and video inputs with text output and a 1M context window on OpenRouter.
Conversation · Reasoning · Tool use
Released 2026-09-03Large model
Qwen
Alibaba multimodal reasoning model for coding assistance, agent workflows, visual understanding, and document or codebase analysis. Accepts text, images, and video with a 1M context window on OpenRouter.
Conversation · Reasoning · Code generation +1
Released 2026-08-26Large model
Qwen
Qwen3.8 27B is an open-weight dense vision-language model from Qwen. It is suited for coding, professional workflows, research, multimodal interaction, and long-running agent tasks, with flexible thinking that can be enabled or disabled.
Conversation · Reasoning · Code generation +3
Released 2026-08-14Medium model
Qwen
Qwen3.8 2.4T A95B is an open-weight sparse mixture-of-experts model from Qwen and the open-weight variant of Qwen3.8 Max, with 95 billion active parameters out of 2.4 trillion total. It is suited for coding, research, complex reasoning, and agentic workflows.
Conversation · Reasoning · Code generation +3
Released 2026-08-12Extra large model
Qwen
The top tier of Alibaba's Qwen3.8 series and the general-availability successor to Qwen3.8 Max Preview. A multimodal reasoning model for complex reasoning, visual understanding, coding and agent workflows.
Conversation · Reasoning · Code generation +4
Released 2026-08-03Extra large model
Qwen
Alibaba vision-language reasoning model for multimodal agents, visual coding, search, and computer interaction, including object recognition and spatial understanding.
Conversation · Reasoning · Code generation +1
Released 2026-07-27Large model
Qwen
The largest model in Alibaba's Qwen3.7 series, text in and text out. Built for agent-centric work: coding, office and productivity tasks, and long-horizon autonomous execution. Clear gains on coding and agent behaviour over earlier Qwen generations, with explicit prompt caching for repeated context.
Conversation · Reasoning · Code generation +4
Released 2026-05-22Extra large model
Qwen
The cost-effective tier of Alibaba's Qwen3.7 series, taking text and image. Carries the series' coding, tool use and productivity work, plus a large vision upgrade. The distinguishing trait is interactive multimodal agency: reading screens, driving GUIs, coding from a visual reference, navigating mobile apps.
Conversation · Reasoning · Code generation +4
Released 2026-05-22Large model
Qwen
Qwen3.5 Plus (April 2026) is a large-scale multimodal language model from Alibaba. It accepts text, image, and video input and produces text output, with a 1M token context window. This is an updated version of Qwen3.5 Plus with tiered pricing above 256K tokens.
Conversation · Reasoning · Code generation +4
Released 2026-04-27Extra large model
Qwen
Qwen3.6 Flash is a fast, efficient language model from Alibaba's Qwen 3.6 series. It supports text, image, and video input with a 1M token context window. Tiered pricing kicks in above 256K tokens. Prompt caching is supported, with both explicit cache read and cache creation pricing.
Conversation · Reasoning · Code generation +3
Released 2026-04-27Medium model
Qwen
An open-weight multimodal model from Alibaba Cloud, 35B total with 3B active per token. Hybrid sparse mixture-of-experts mixing Gated DeltaNet linear attention with standard gated attention. 262K native context, extensible to 1M with YaRN. Text, image and video. Thinking mode built in. Apache 2.0.
Conversation · Reasoning · Code generation +3
Released 2026-04-27Medium model
Qwen
A dense 27B from Alibaba's Qwen team, released April 2026. Text, image and video across a 262,144 token context. Built for agentic coding and reasoning, clearest on repository-level code comprehension and front-end work. Thinking mode built in and persistent. 201 languages, Apache 2.0.
Conversation · Reasoning · Code generation +4
Released 2026-04-27Medium model
Qwen
Alibaba Cloud's proprietary Qwen3.6 Max, a sparse mixture of experts at roughly 1 trillion total parameters. For agentic coding, tool use and long-context reasoning over a 262K window, with a thinking mode that keeps traces across turns. Model Studio and Qwen Studio only, no open weights.
Conversation · Reasoning · Code generation +4
Released 2026-04-27Extra large model
Qwen
The preview of Qwen 3.6 Plus, on a hybrid architecture that improves efficiency and scaling over the 3.5 series, with more reliable agent behaviour. Strongest on agentic coding and front-end development. Free tier on OpenRouter, which collects prompt and completion data.
Conversation · Reasoning · Code generation +1
Released 2026-03-30Large model
Qwen
A multimodal foundation model in the Qwen3.5 family at 9B parameters. Unified vision-language design with early fusion of multimodal tokens: text, image and video in, text out, with reasoning built in.
Conversation · Reasoning · Code generation +1
Released 2026-03-10Small model
Qwen
The Qwen3.5 Series 35B-A3B is a native vision-language model designed with a hybrid architecture that integrates linear attention mechanisms and a sparse mixture-of-experts model, achieving higher inference efficiency. Its overall performance is comparable to that of the Qwen3.5-27B.
Conversation · Reasoning · Code generation +4
Released 2026-02-25Medium model
Qwen
The Qwen3.5 27B native vision-language Dense model incorporates a linear attention mechanism, delivering fast response times while balancing inference speed and performance. Its overall capabilities are comparable to those of the Qwen3.5-122B-A10B.
Conversation · Reasoning · Code generation +4
Released 2026-02-25Medium model
Qwen
The 122B-A10B vision-language model in the Qwen3.5 series, on a hybrid of linear attention and a sparse mixture of experts. Second only to Qwen3.5-397B-A17B overall, with text well ahead of Qwen3-235B-2507 and vision ahead of Qwen3-VL-235B.
Conversation · Reasoning · Code generation +4
Released 2026-02-25Large model
Qwen
The Flash tier of Qwen3.5, a native vision-language model on a hybrid of linear attention and a sparse mixture of experts. Built for fast responses, and a clear step up from the Qwen3 series on both text and multimodal tasks.
Conversation · Reasoning · Code generation +4
Released 2026-02-25Medium model
Qwen
The Plus tier of Qwen3.5, a native vision-language model on a hybrid of linear attention and sparse mixture of experts for cheaper inference. Alibaba reports it level with leading models, and a clear step up from Qwen3 on text and multimodal work. Text, image and video in, with reasoning and tool use.
Conversation · Reasoning · Code generation +4
Released 2026-02-16Extra large model
Qwen
The 397B-A17B vision-language model in the Qwen3.5 series, on a hybrid of linear attention and sparse mixture of experts. Alibaba reports it level with leading models across language, logic, code, agent tasks, image and video understanding and GUI interaction. 201 languages and dialects.
Conversation · Reasoning · Code generation +4
Released 2026-02-16Extra large model
Qwen
The reasoning tier of Qwen3-Max, scaled on both model capacity and reinforcement learning compute. Heavy Mode iterates on an answer at test time, tools include search and a code interpreter, and it can switch between normal and compute-heavy reasoning mid-conversation.
Conversation · Reasoning · Code generation +2
Released 2026-02-09Extra large model
Qwen
An open-weight model for coding agents and local development. Sparse MoE with 80B total parameters and 3B activated per token, performing like models with 10 to 20x the active compute. Non-thinking mode only.
Conversation · Code generation · Agentic tool use +1
Released 2026-02-04Large model
Qwen
An image generation foundation model in the Qwen series that achieves significant advances in complex text rendering.
Image generation
Released 2026-01-20Large model
Qwen
A fast Qwen text-to-image model optimized by PrunaAI for speed on Replicate.
Image generation
Released 2025-12-11Large model
Qwen
The larger of Alibaba's two hosted Qwen3 Coder tiers, sold through Model Studio rather than released as open weights. 128K context, with OpenAI-compatible and Anthropic-compatible endpoints.
Conversation · Reasoning · Code generation +1
Released 2025-09-17Large model
Qwen
The cheap tier of Alibaba's hosted Qwen3 Coder line, aimed at autocomplete and quick edits where latency matters more than depth. 128K context, Model Studio only, no open weights.
Conversation · Reasoning · Code generation +1
Released 2025-09-17Medium model
Qwen
The instruct member of Qwen3-Next, tuned for fast, stable answers with no visible chain of thought. It prioritises throughput and consistency on very long inputs and multi-turn dialogue, which is what makes it usable for retrieval-augmented generation and tool-calling agents.
Conversation · Reasoning · Code generation +1
Released 2025-09-11Large model
Qwen
The reasoning member of Qwen3-Next, emitting thinking traces by default in thinking-only mode. Built for maths proofs, code synthesis and debugging, logic and agent planning, and tuned to stay stable across long chains without drifting off task.
Conversation · Reasoning · Code generation +1
Released 2025-09-11Large model
Qwen
Qwen Plus 0728 with reasoning switched on. The same Qwen3-based 1 million token hybrid model, spending extra tokens to think before it answers.
Conversation · Reasoning · Code generation +1
Released 2025-09-08Large model
Qwen
Qwen Plus 0728, built on Qwen3: a hybrid reasoning model with a 1 million token context, balancing quality, speed and cost.
Conversation · Reasoning · Code generation +1
Released 2025-09-08Large model
Qwen
Alibaba's largest Qwen3 model, an update on the January 2025 release with better mathematics, coding, logic and science, and fewer hallucinations on open-ended questions. Over 100 languages, tuned for retrieval-augmented generation and tool calling. No dedicated thinking mode.
Conversation · Reasoning · Code generation +3
Released 2025-09-05Extra large model
Qwen
The thinking variant of Qwen3-30B-A3B, a 30B mixture-of-experts model that keeps its reasoning trace separate from the final answer. Longer output budgets than earlier 30B releases, with gains across logic, mathematics, science, coding and multilingual benchmarks.
Conversation · Reasoning · Code generation +1
Released 2025-08-29Large model
Qwen
A 30.5B mixture-of-experts model from Qwen with 3.3B active parameters per inference, running in non-thinking mode only. It answers directly rather than reasoning aloud, and beats the non-instruct variant on open-ended and subjective tasks while holding its coding and factual scores.
Conversation · Reasoning · Code generation +1
Released 2025-07-29Large model
Qwen
The thinking-only variant of Qwen3-235B-A22B, activating 22B of 235B parameters and holding 262,144 tokens of context. It always emits a reasoning trace and is built for long outputs, up to 81,920 tokens, on maths, science and long-form generation.
Conversation · Reasoning · Code generation +2
Released 2025-07-25Extra large model
Qwen
Qwen's 480B mixture-of-experts coding model, activating 35B parameters per pass from 8 of 160 experts. Built for agentic coding: function calling, tool use, and reasoning over a whole repository rather than a single file.
Conversation · Reasoning · Code generation +3
Released 2025-07-23Extra large model
Qwen
The July 2025 instruct refresh of Qwen3-235B-A22B, activating 22B of 235B parameters per pass. Native 262K context and no thinking mode: it answers directly. Gains over the base variant are largest in knowledge coverage, long-context reasoning, coding and multilingual maths.
Conversation · Reasoning · Code generation +1
Released 2025-07-21Extra large model
Qwen
A 0.6B dense model from the Qwen3 family, small enough to run on a laptop. Switches between thinking mode for hard tasks and non-thinking mode for chat. Trained on 36 trillion tokens across 119 languages, with tool use and multilingual support.
Conversation · Code generation
Released 2025-04-29Small model
Qwen
Qwen3 at 30.5B parameters with 3.3B activated. Reasoning, multilingual work and agent tasks, with a thinking/non-thinking mode switch. Up to 131K context with YaRN. Free tier on OpenRouter.
Conversation · Reasoning · Code generation +1
Released 2025-04-28Large model
Qwen
A 235B mixture-of-experts model from Alibaba's Qwen team, activating 22B parameters per forward pass. Switches between a thinking mode for hard tasks and a non-thinking mode for chat. Strong on reasoning, tool calling and over 100 languages. 32K context, extendable to 131K.
Conversation · Reasoning · Code generation +1
Released 2025-04-28Extra large model
Qwen
Qwen's 32B reasoning model. It thinks before answering, which buys it a wide margin over conventional instruction-tuned models on hard problems and puts it in range of much larger reasoning models like DeepSeek-R1 and o1-mini.
Conversation · Reasoning · Code generation +1
Released 2025-03-05Large model
Recraft
The speed and cost tier of Recraft V4.1 for text-to-image generation. Produces raster images at roughly 1K resolution across several aspect ratios, priced at $0.007 per image on the listed OpenRouter endpoint.
Image generation
Released 2026-09-23Large model
Recraft
Recraft V4.1 refines V4's design taste with quieter, more natural photorealism, dreamier gradients, and sharper illustration styles. The model still makes intentional decisions about composition, lighting, and material realism while rendering accurate text from prompts up to 10,000 characters.
Image generation
Released 2026-05-14Large model
Recraft
Recraft V4.1 Pro pairs V4.1's refined photorealism and sharper illustration with maximum-quality, high-resolution output. Built for final production assets, print, and large-format work where detail and fidelity matter most.
Image generation
Released 2026-05-14Large model
Recraft
A ground-up rebuild of Recraft's image generation with design judgment built in: the model makes deliberate choices about composition, lighting, colour and material realism. Takes prompts up to 10,000 characters and renders text accurately.
Image generation
Released 2026-02-17Large model
Recraft
Recraft V4 Pro renders at roughly 2048px, scalable to 3072px, with the same design taste as V4. Built for print, large-format displays and detailed professional work.
Image generation
Released 2026-02-17Large model
Recraft
Recraft V4 SVG generates production-ready SVG vector images from text prompts with real paths, structured layers, and clean geometry. The only image model that produces native, editable vector output.
Image generation
Released 2026-02-17Large model
Recraft
Recraft V4 Pro SVG combines V4 Pro's design taste with detailed geometric vector output. Produces detailed SVG files with finer paths, clean layers, and editable output suitable for professional design systems.
Image generation
Released 2026-02-17Large model
Recraft
Recraft V3 (code-named red_panda) is a SOTA text-to-image model ranked #1 on Artificial Analysis benchmarks. The first model to offer image generation with text of any size and length, supporting both raster and vector output formats across a wide list of styles.
Image generation
Released 2024-10-29Large model
Recraft
Recraft V3 SVG generates high-quality vector SVG images from text prompts, including logotypes, icons, and illustrations. Produces real editable SVG files with clean paths and structured layers.
Image generation
Released 2024-10-29Large model
Sourceful
The highest quality variant of Sourceful's Riverflow 2.5, for work where control matters more than speed. The series treats image generation as a production workflow: a reasoning model plans multi-step edits and judges candidates before accepting one. Reasoning effort is set on the request.
Image generation
Released 2026-06-04Large model
Sourceful
The speed-optimized variant of Sourceful's Riverflow 2.5, for production deployments with a latency budget. Like the rest of the series it plans multi-step edits with a reasoning model and judges candidates before accepting one, with reasoning effort set on the request.
Image generation
Released 2026-06-04Large model
Sourceful
The top variant of Sourceful's Riverflow 2.0 line, for fine control and text rendering. The series runs an integrated reasoning model over image generation and editing.
Image generation
Released 2026-02-02Large model
Sourceful
The fastest variant of Sourceful's Riverflow 2.0 line, for production and latency-critical work. The series runs an integrated reasoning model over image generation and editing.
Image generation
Released 2026-02-02Large model
Stability AI
Stability AI's latest audio generation model for music, sound effects, and ambient audio up to 3 minutes long.
Music generation · Audio generation
Released 2025-04-20Large model
Stability AI
Stable Diffusion 3.5 Medium balances quality and speed, offering modern diffusion performance with broad aspect ratio support.
Image generation
Released 2024-10-29Large model
Stability AI
Stable Diffusion XL (SDXL), a widely used open text-to-image diffusion model known for versatility and community tooling.
Image generation
Released 2023-07-26Large model
Tencent
Tencent's 770B-parameter mixture-of-experts model activates 49B parameters per token. It is built for coding agents and tool-heavy workflows that need planning, context continuity and sustained execution across a 1M-token window.
Conversation · Reasoning · Code generation +5
Released 2026-08-28Extra large model
Tencent
Tencent mixture-of-experts model with 295B total and 21B active parameters, using 192 experts with top-8 routing. Designed for reasoning and agent workflows with configurable reasoning effort.
Conversation · Reasoning · Tool use
Released 2026-07-06Extra large model
Tencent
Tencent's native multimodal image model, tuned by PrunaAI for speed. Strong prompt understanding across a range of visual styles.
Image generation
Released 2025-10-01Large model
Upstage
Upstage's compact mixture-of-experts model with 35 billion total parameters and 3 billion active parameters. It supports text generation, tool calling, and a 524,288-token context.
Conversation · Reasoning · Analysis +2
Released 2026-09-23Small model
Upstage
Upstage language model for long-horizon agent workflows, office productivity, and document-intensive tasks. Supports text input and output with a 524K context window on OpenRouter.
Conversation · Reasoning · Tool use
Released 2026-08-10Large model
Upstage
Upstage's Mixture-of-Experts model: 102B total parameters, 12B active per forward pass. Tuned for Korean, with English and Japanese support. Deprecating 2 March 2026.
Conversation · Reasoning · Code generation +1
Released 2026-01-27Large model
Xiaomi
Xiaomi MiMo-V2.6 Flash accepts text, image, audio, and video inputs and produces text. Its OpenRouter endpoint supports a 1M context window with lower token pricing than the Pro variant.
Conversation · Reasoning · Tool use
Released 2026-09-21Large model
Xiaomi
Xiaomi MiMo-V2.6 Pro accepts text, image, audio, and video inputs and produces text, with a 1M context window on OpenRouter.
Conversation · Reasoning · Tool use
Released 2026-09-21Large model
Xiaomi
Xiaomi's native omnimodal MiMo-V2.5, roughly half the inference cost of the Pro tier and ahead of MiMo-V2-Omni on image and video understanding. Its 1M context takes whole documents, long conversations and complex task state in one pass, which is what makes it usable inside agent frameworks.
Conversation · Reasoning · Code generation +2
Released 2026-04-22Large model
Xiaomi
Xiaomi's largest MiMo-V2.5, ranked near the top on ClawEval, GDPVal and SWE-bench Pro. It runs professional tasks that would take a human expert days, chaining more than a thousand tool calls without supervision, across a context of up to 1M tokens.
Conversation · Reasoning · Code generation +3
Released 2026-04-22Extra large model
Xiaomi
Xiaomi's largest MiMo model, over 1T total parameters with a 1M context, tuned for agent frameworks like OpenClaw. Xiaomi places it in the top tier on PinchBench and ClawBench, near Opus 4.6. The intended job is orchestrating workflows and production engineering tasks rather than answering single questions.
Conversation · Reasoning · Code generation +3
Released 2026-03-18Extra large model
Xiaomi
An omni-modal model that takes image, video and audio in one architecture. Adds visual grounding, multi-step planning, tool use and code execution for tasks that span modalities.
Conversation · Reasoning · Code generation +2
Released 2026-03-18Large model
Xiaomi
Xiaomi's open-source MiMo-V2-Flash, a mixture-of-experts with 309B total and 15B active on a hybrid attention design. 256K context, thinking toggle. Xiaomi reports it top open-source on SWE-bench Verified and SWE-bench Multilingual, near Claude Sonnet 4.5 at about 3.5% of the cost.
Conversation · Reasoning · Code generation +2
Released 2025-12-14Extra large model
Z.ai
A faster serving variant of GLM 5.3 with the same capabilities. It accepts text, supports tool calling and reasoning, and provides a one-million-token context.
Conversation · Reasoning · Analysis +2
Released 2026-09-23Large model
Z.ai
GLM-5.1's gain over its predecessor is duration. Where earlier models were built around minute-long interactions, it works a single task for more than 8 hours, planning, executing and correcting itself until it has something engineering-grade to hand back.
Conversation · Reasoning · Code generation +4
Released 2026-04-07Large model
Z.ai
Z.ai's GLM-5 Turbo, for fast inference in agent-driven environments such as OpenClaw. Tuned for long execution chains: instruction decomposition, tool use, scheduled and persistent execution, and stability across extended tasks.
Conversation · Reasoning · Code generation +5
Released 2026-03-15Medium model
Zhipu AI
Z.ai GLM 5.3 FlashX accepts text, image, and video inputs and produces text, with a 1M context window on OpenRouter.
Conversation · Reasoning · Tool use
Released 2026-09-18Large model
Zhipu AI
Z.ai multimodal model for efficient coding and long-horizon agent tasks, using hybrid sparse and linear attention. Accepts text, image, and video inputs with a 1,310,720-token context window on OpenRouter.
Conversation · Reasoning · Code generation +1
Released 2026-08-26Large model
Zhipu AI
Z.ai reasoning model for complex software engineering and long-horizon agent tasks. Supports text input and output with a 1,310,720-token context window on OpenRouter.
Conversation · Reasoning · Code generation +1
Released 2026-08-18Large model
Zhipu AI
Z.ai's GLM-5.2, built for long-horizon work. A genuinely usable 1M-token context window lets it hold project-level engineering context, run long tasks, follow engineering standards, and carry a project from requirements to multi-platform deployment in one task.
Conversation · Reasoning · Code generation +4
Released 2026-06-16Extra large model
Zhipu AI
Z.ai's largest open-source model, built for systems design and agent runs that last. It plans, executes and self-corrects across large programming tasks rather than emitting a function and stopping, and Z.ai positions it against leading closed models on production code.
Conversation · Reasoning · Code generation +3
Released 2026-02-11Extra large model
Zhipu AI
A 30B-class GLM tuned for agentic coding: coding strength, long-horizon task planning and tool collaboration. Z.ai reports top results among open-weight models of the same size on several public benchmarks.
Conversation · Reasoning · Code generation +1
Released 2026-01-27Medium model
Zhipu AI
GLM 4.6 expands the GLM family with a 200K-token context window, stronger coding benchmarks, and more reliable multi-step reasoning. It integrates deeply with agent frameworks to orchestrate tool use and produces more natural writing for long-form chat.
Conversation · Reasoning · Code generation +2
Released 2025-09-30Extra large model
Zhipu AI
Z.ai's GLM-4.5, a mixture-of-experts model built for agent applications with 128K context. It runs in two modes: thinking, for multi-step reasoning and tool use, and non-thinking, for instant answers.
Conversation · Reasoning · Code generation +1
Released 2025-07-25Large model
Zhipu AI
The compact GLM-4.5, same mixture-of-experts design at a smaller parameter count. Keeps the two-mode split, thinking for reasoning and tool use, non-thinking for real-time chat, toggled by a boolean on the request.
Conversation · Reasoning · Analysis
Released 2025-07-25Large model
Zhipu AI
A cost-effective 32B foundation model with much stronger tool use, web search and code work than GLM 3. From the lab behind the THUDM models.
Conversation · Reasoning · Code generation +2
Released 2025-07-24Large model
xAI
Grok 4.7 supports text, image, and file inputs with text output and a 500K context window on OpenRouter.
Conversation · Reasoning · Tool use
Released 2026-09-21Large model
xAI
SpaceXAI's Grok 4.6, aimed at coding, knowledge work and STEM. 500K-token context window, with text, image and file input, and reasoning enabled via the API.
Conversation · Reasoning · Code generation +3
Released 2026-08-12Large model
xAI
xAI's Grok 4.5, aimed at coding, knowledge work and STEM. 500K-token context window, text and image in, reasoning enabled via the API.
Conversation · Reasoning · Code generation +1
Released 2026-07-08Large model
xAI
Grok 4.3 is xAI's reasoning model with a 1M token context window. It accepts text and image inputs with text output, suited for agentic workflows, instruction-following tasks, and applications requiring high factual accuracy. Supports parallel tool calling, structured outputs, and exposed reasoning.
Conversation · Reasoning · Code generation +1
Released 2026-05-02Large model
xAI
xAI's Grok 4.20 Beta, built for speed and agentic tool calling. xAI claims the lowest hallucination rate on the market and strict prompt adherence. Reasoning toggles on and off via the API.
Conversation · Reasoning · Code generation +4
Released 2026-03-12Extra large model
xAI
A Grok 4.20 variant for agent-based workflows. Multiple agents run in parallel to research, coordinate tool use and synthesise across complex tasks. Reasoning effort sets agent count: low and medium use 4, high and xhigh use 16.
Conversation · Reasoning · Code generation +5
Released 2026-03-12Extra large model
xAI
xAI's image generation model, pairing Grok's prompt understanding with high-quality output. Aimed at detailed, creative images that follow the prompt closely.
Image generation
Released 2025-12-01Large model
xAI
xAI's fast tier of Grok 4.1, shipped alongside the Agent Tools API for search and code execution. Sold on the xAI API and Oracle Cloud, in reasoning and non-reasoning variants.
Conversation · Reasoning · Code generation +2
Released 2025-11-21Extra large model
xAI
Grok 4 Fast is xAI's latest multimodal model with SOTA cost-efficiency and a 2M token context window. It comes in two flavors: non-reasoning and reasoning. Reasoning can be enabled via the API.
Conversation · Analysis · Code generation
Released 2025-09-19Large model
xAI
xAI's cheap, fast coding model, built for agents that make many small edits in an editor rather than one long deliberation. 256K context, available through the xAI API and Azure AI Foundry.
Conversation · Reasoning · Code generation +1
Released 2025-08-26Medium model
xAI
Grok 4 is xAI's latest reasoning model with a 256k context window. It supports parallel tool calling, structured outputs, and both image and text inputs. Note that reasoning is not exposed, reasoning cannot be disabled, and the reasoning effort cannot be specified.
Conversation · Reasoning · Code generation +1
Released 2025-07-09Large model
xAI
A small thinking model for reasoning-heavy work that does not need deep domain knowledge. Strongest on math and quantitative tasks. The thinking traces are visible.
Conversation · Reasoning · Analysis
Released 2025-04-09Small model
xAI
xAI's top Grok 3 tier, aimed at data extraction, coding and summarization, with domain knowledge in finance, healthcare, law and science. Beats Grok 3 Mini on high-thinking tasks.
Conversation · Reasoning · Code generation +2
Released 2025-04-09Extra large model
xAI
Grok 3 Thinking exposes the full chain-of-thought process during problem-solving, including error backtracking and alternative solution exploration. Scores 84.6% on GPQA Diamond benchmark for expert-level Q&A.
Conversation · Reasoning · Code generation +1
Released 2025-02-19Extra large model
xAI
xAI's Grok 3, trained on the Colossus cluster, with a Big Brain Mode that spends extra compute on hard problems. Scored 1402 Elo on LMArena and 93.3% on AIME 2025.
Conversation · Reasoning · Code generation +1
Released 2025-02-18Extra large model
351+ models you haven't met. $0 to continue. Seriously.
Free account. No card required. By continuing, you agree to Rival's Terms and Privacy policy
