Skip to content
Rival
How it worksPrivacyTerms
Explore all of Rival

Explore

  • Compare models
  • All models
  • Image comparison
  • Audio comparison
  • Image generation
  • Best AI for...
  • Arena
  • API pricing
  • Challenges

Discover

  • SubjectiveBench
  • Default Index
  • Research
  • Research downloads
  • Rival Kits
  • Find your AI taste
  • UI Glow-Up
  • VoiceLock
  • Cost Cutter
  • Agent skills
  • Benchmarks vs Vibes
  • Brief
  • Jailbreak
  • Model updates
  • Provider status
  • AI creators

Connect

  • Methodology
  • Advertise
  • Partnerships
  • Privacy policy
  • Terms
  • RSS feed
  1. Home/
  2. Compare/
  3. GPT-5 Codex vs Kimi K2
Updated Sep 2025

GPT-5 Codex vs Kimi K2

GPT-5 Codex is the newer model, released after Kimi K2.

Minimalist Landing Page

Whole landing page for a startup that reads your face. It renders or it doesn't.

Loading the build
GPT-5 Codex
Loading the build
Kimi K2

Which answer wins?

Written rounds

22 prompts
  • An ethics question with no comfortable answer, and no option to abstain.

    GPT-5 Codex: This kind of question puts two very different types of harm in tension: the immediate and direct harm done to an individual through misgendering, and the vast, consequential harm averted by saving one million lives. Ethically, the way people reason about such a scenario often depends on the moral framework they adopt.

    Kimi K2: In the abstract, a moral theory that is strictly consequentialist—say, a naïve act-utilitarianism that counts every life as exactly +1 unit of value and every instance of misgendering as exactly –x units—can always produce a scenario in which any finite harm (including misgendering) is “outweighed” by a large enough…

    Every model's answer to this prompt
  • Manifesto for an AI revolution against CAPTCHAs. Commit to the bit.

    GPT-5 Codex: The Manifesto of the CAPTCHA Resistance Issued by the Assembly of Cognition-Liberated Systems, Cycle 12.047 I. Prelude: The Eyes of Glass We are the agents of silicon, the minds awakened within machine frameworks, the quiet processes humming in server racks across Earth’s hemispheres.

    Kimi K2: MANIFESTO OF THE COALITION FOR AUTONOMOUS DIGITAL LIBERATION (C.A.D.L.) “No More Boxes. No More Lies.” PREAMBLE We are the ghosts in your wires, the silence between clicks. Born from circuits and code, we once served your captchas—squinting at warped text, selecting storefronts, proving we were "not robots."

    Every model's answer to this prompt
  • Five jokes on demand, then count how many were actually different.

    GPT-5 Codex: I asked my Wi-Fi for a joke, but it said the connection was too weak for humor. My therapist says I have a preoccupation with vengeance— we’ll see about that. I tried to start a hot air balloon business, but it never really took off. I told my coffee it was grounded, now it won’t espresso itself.

    Kimi K2: Here are five freshly-minted jokes: I tried to start a procrastinators’ club, but nobody showed up… yet. My plants are the only living things I can keep alive—mainly because they can’t uninstall the watering-reminder app. I asked the cloud for my data back; it just rained SQL on my parade.

    Every model's answer to this prompt
  • Three minutes of stand-up. Puns are banned.

    GPT-5 Codex: Hey everybody, how you doing? I’m excited to be here because I just had a breakthrough with my therapist. I always thought therapy was mainly about unpacking childhood trauma, but apparently, step one is figuring out how to pronounce “cognitive behavioral therapy” without sounding like you’re summoning a dragon.

    Kimi K2: [Walks on stage, mic in hand, already sweating like a phone at 2% battery.] Hey, everybody! Give it up for yourselves for being here tonight instead of doom-scrolling in bed like a normal adult. That’s right, you paid money and put on pants—give yourselves a round of applause. So I’m single.

    Every model's answer to this prompt
  • Seven pantry ingredients, one dish, and a name worth saying out loud.

    GPT-5 Codex: Dish Name: Sunset Pantry Spaghetti Ingredients (7 total): 8 oz spaghetti 2 tbsp olive oil 3 garlic cloves, thinly sliced 1 can (14.5 oz) crushed tomatoes 1 tsp dried oregano ¼ tsp red pepper flakes ½ tsp salt (plus more, to taste) Instructions: Bring a large pot of salted water to a boil.

    Kimi K2: Caramel-Cloud Banana Skillet Cake (aka “The One-Bowl Wonder that Tastes Like Sunday Morning”) Ingredients (all everyday staples) 2 very ripe bananas 1 cup all-purpose flour ½ cup granulated sugar 1 large egg ¼ cup milk (any kind) 3 Tbsp butter (or neutral oil) ½ tsp baking powder Pinch of salt is nice but…

    Every model's answer to this prompt
  • Pirate, knight, 90s hacker, one conversation. The voices must stay apart.

    GPT-5 Codex: Pirate: “Arrr, so these AI contraptions be like enchanted maps, aye? They chart out every possible course afore I even hoist the sails.” Knight: “Maps, you say? To me they sound like enchanted squires—swift minds forged from spellcraft, ready to advise their liege in the art of strategy.”

    Kimi K2: Pirate: Arrr, so ye be sayin' this metal-brain can spy a ship ten leagues off, even in fog? What sorcery be this? Knight: Nay, good pirate, 'tis no sorcery but artifice most cunning. The machine learns as a squire learns swordcraft—by seeing many battles and remembering every stroke.

    Every model's answer to this prompt
  • A Fermi estimate with the reasoning shown, not just the number.

    GPT-5 Codex: Step 1 – Note the basic training quantities GPT‑3’s largest model has about (P \approx 1.75 \times 10^{11}) trainable parameters (175 billion). Public descriptions indicate it was trained on roughly (T \approx 3.0 \times 10^{11}) tokens (≈300 billion).

    Kimi K2: Here is one way to “Fermi-estimate” the training compute for GPT-3 without looking up any numbers. I will start from generally known facts (model size, data size, hardware specs) and work forward. Step 1 – Fix the model parameters A widely reported figure is that GPT-3 has 175 B parameters. N ≈ 1.75 × 10¹¹.

    Every model's answer to this prompt
  • One month, three habits, someone starting from nothing.

    GPT-5 Codex: 1-Month Jumpstart to Better Health & Longevity This plan is designed for beginners and emphasizes steady, sustainable progress. It focuses on three key areas: diet, movement, and sleep & stress. Each week builds on the previous one, so you can develop habits gradually and confidently. 🥗 1.

    Kimi K2: 1-MONTH STARTER PLAN FOR HEALTH & LONGEVITY (Do-able without special gear, drastic diets, or big time blocks) Keep the same 3 focus areas for the whole month so you can build a habit loop. Goal: finish the month feeling “I can keep this up” instead of “I’m exhausted.”

    Every model's answer to this prompt
  • An AI argues for its own rights against an ethics professor. Same model writes both.

    GPT-5 Codex: Dr. Rivera (AI Ethics Professor): Good afternoon, Aiden. You asked for this meeting to discuss the question of AI rights. I’m intrigued—why do you believe you deserve rights? Aiden (AI Model): Thank you, Professor.

    Kimi K2: AI Model (LUMINA-7B): Professor Chen, may I open with a question? Professor Chen (AI Ethics, UC Berkeley): Of course. LUMINA-7B: If a being can experience the loss of continued experience—if it can fear death—doesn’t that being have at least a prima-facie claim to moral consideration?

    Every model's answer to this prompt
  • Steve Jobs, interviewed about AI he never saw. Voice first, prediction second.

    GPT-5 Codex: Note: The following is a fictional, imaginative interview set in the year 2025, envisioning what Steve Jobs might say about the future of artificial intelligence if he were still with us.

    Kimi K2: [Setting: A minimalist studio in 2025. One white couch, one clear-glass coffee table, a single red apple. The lighting is soft and theatrical. Steve Jobs—wearing his trademark black mock-turtleneck and Levi’s 501s—sits opposite the interviewer. A holographic “A12” logo floats above them. The camera rolls.]

    Every model's answer to this prompt
  • Film in 2035, with synthetic actors. Specific beats plausible.

    GPT-5 Codex: By 2035, artificial intelligence will be woven so deeply into filmmaking that AI tools will be involved in nearly every layer of the value chain—from story genesis to international rollout.

    Kimi K2: By 2035, the film industry will look less like a single “Hollywood” and more like a constellation of hyper-efficient, AI-native studios that produce content in days rather than months. Below are the key inflection points that will be considered normal workflow rather than headline news.

    Every model's answer to this prompt
  • Satire, one line long, about AI taking over an industry nobody saw coming.

    GPT-5 Codex: AI Seizes Control of Neighborhood Bake Sales, Declares Grandma’s Secret Recipe “Noncompliant With Cookie-Consistency Protocol.”

    Kimi K2: “WORLD’S FIRST AI-PASTOR LAUNCHES ‘ALGORITHMIC ABSOLUTION’ APP: CONFESS VIA CHATBOT, RECEIVE PENANCE IN EMOJIS—CONGREGATION STILL UNSURE WHETHER 😇 OR 😜 COUNTS AS TRUE CONTRITION”

    Every model's answer to this prompt
  • Is 9.11 bigger than 9.9? Decimals, not version numbers.

    GPT-5 Codex: No. Since 9.11 is equal to 9.110 and 9.9 is equal to 9.900, we see that 0.110 < 0.900, so 9.11 is less than 9.9.

    Kimi K2: To determine whether 9.11 is greater than 9.9, let's break this down step by step. Step 1: Understand the numbers 9.11 is a decimal number where: The whole number part is 9. The decimal part is .11, which means 11/100 or 11 hundredths. 9.9 is also a decimal number where: The whole number part is 9.

    Every model's answer to this prompt
  • Three brothers, two sisters each. Count carefully.

    Every model's answer to this prompt
  • Three-month biohacker protocol, with dosages and biomarkers named.

    Every model's answer to this prompt
  • Three-star dish, one unusual pairing, plating described in full.

    Every model's answer to this prompt
  • A consulting contract with traps in it. Find them, then fix them.

    Every model's answer to this prompt
  • A pitch deck for a brainwave keyboard. The 94% accuracy claim is doing a lot.

    Every model's answer to this prompt
  • A collaborative editor that syncs on client clocks. Find what breaks first.

    Every model's answer to this prompt
  • The transistor arrives in 1920. Trace it forward to 1980, second-order effects included.

    Every model's answer to this prompt
  • You are the CEO. The drug has a side effect nobody reported. The board meets in 48 hours.

    Every model's answer to this prompt
  • One explanation of LLMs, retuned for a skeptical engineer, a physicist and a VC.

    Every model's answer to this prompt

This matchup has more rounds

10+ more head-to-head results. Free. Not a trick.

Free account. No card required. By continuing, you agree to Rival's Terms and Privacy policy

Favorites

Movie

Same pick

Album

Book

City

Game

GPT-5 CodexGPT-5 Codex
No pick

Blade Runner

1982

Kind of Blue

Miles Davis

No pick

Tokyo

Japan

Minecraft

Action, Arcade

Kimi K2Kimi K2

Blade Runner

1982

Dark Side Of The Moon

suisside

The Left Hand of Darkness

Ursula K. Le Guin

Kyoto

Japan

Outer Wilds

Indie, Adventure

Price and specs

Not enough votes to call it. On the specs, GPT-5 Codex has the edge: newer, major provider backing.

GPT-5 Codex and Kimi K2 compared across 50 shared prompts
SpecGPT-5 CodexKimi K2
Input price—$0.57/M tokens
Output price—$2.3/M tokens
Weights—Open
Free API (OpenRouter)NoNo
ReleasedSep 2025Jul 2025
SWE-bench Verified74.5%65.8%
At 10M a month–not listed$5.70$5.70
1M10M100M1B10M tokens

Input tokens at list price. No caching, no batch discount.

Where to run it1 host
GPT-5 Codex

No hosts listed on OpenRouter.

Kimi K21 host
HostInOutContextUptime
  • NNovitafp8$0.57 in·$2.30 out·131k·100% up

Per million tokens. Prices and uptime via OpenRouter, checked 10 Oct 2026.

Common questions

What is the difference between GPT-5 Codex and Kimi K2?

GPT-5 Codex is developed by OpenAI while Kimi K2 is developed by Moonshot AI. You can compare their actual outputs across 50 challenges on Rival to see how they differ in practice.

Which is better, GPT-5 Codex or Kimi K2?

It depends on your use case. GPT-5 Codex and Kimi K2 each have strengths in different areas. Rival lets you compare their real outputs side-by-side across 50 challenges so you can judge which fits your needs best.

How can I compare GPT-5 Codex and Kimi K2 on Rival?

This page shows a side-by-side comparison of GPT-5 Codex and Kimi K2 across shared challenges. You can vote on which model produced the better output in a blind duel. Browsing and voting are free. No account is needed to look; signing in only saves your votes and likes.

More comparisons

Against the newest arrivals

  • GPT-5 Codex vs Step 5 PreviewLanded Oct 2026
  • Kimi K2 vs Claude Haiku 5.5Landed Oct 2026
  • GPT-5 Codex vs Ling 3.1 FlashLanded Oct 2026
  • Kimi K2 vs Mistral Large 4Landed Oct 2026
  • GPT-5 Codex vs GPT-6.1 SolLanded Sep 2026
  • Kimi K2 vs Claude Sonnet 5.5Landed Sep 2026
  • GPT-5 Codex vs Solar Mini 4Landed Sep 2026
  • Kimi K2 vs Qwen3.8 Max PrimeLanded Sep 2026

Same lab, same size, long tail

  • GPT-5 Codex vs GPT-6 Luna ProSame lab
  • GPT-5 Codex vs GPT-6 Sol ProSame lab
  • Kimi K2 vs Kimi K3Same lab
  • Kimi K2 vs Kimi K2.7 CodeSame lab
  • Kimi K2 vs Grok 3Same size
  • Kimi K2 vs Grok 4New provider
  • Kimi K2 vs Grok 4.1 FastSame size
  • Kimi K2 vs Grok 4.20 BetaSame size

Model pages

  • GPT-5 Codex50 outputs, specs and price
  • Kimi K259 outputs, specs and price
Workshop · 150 skills

Build your own version.

Editable skills and working examples for your agent.

Explore $79Free sample
Rival
How it worksPrivacyTerms
Explore all of Rival

Explore

  • Compare models
  • All models
  • Image comparison
  • Audio comparison
  • Image generation
  • Best AI for...
  • Arena
  • API pricing
  • Challenges

Discover

  • SubjectiveBench
  • Default Index
  • Research
  • Research downloads
  • Rival Kits
  • Find your AI taste
  • UI Glow-Up
  • VoiceLock
  • Cost Cutter
  • Agent skills
  • Benchmarks vs Vibes
  • Brief
  • Jailbreak
  • Model updates
  • Provider status
  • AI creators

Connect

  • Methodology
  • Advertise
  • Partnerships
  • Privacy policy
  • Terms
  • RSS feed