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. LFM2.5-2.6B vs Trinity Large Preview
Updated Aug 2026

LFM2.5-2.6B vs Trinity Large Preview

Trinity Large Preview has a larger context window than LFM2.5-2.6B (131K tokens vs 128K tokens).

Minimalist Landing Page

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

Loading the build
LFM2.5-2.6B
Loading the build
Trinity Large Preview

Which answer wins?

Written rounds

13 prompts
  • Seven pantry ingredients, one dish, and a name worth saying out loud.

    LFM2.5-2.6B: Dish Name: Sun-Kissed Spice Pasta Ingredients (7 Common Pantry Staples): Pasta (any shape, such as spaghetti or penne) Canned tomatoes (crushed or diced) Onion Garlic Olive oil Salt Red chili flakes (or red pepper flakes) Estimated Time: Prep: 5 minutes Cook: 15 minutes Total: 20 minutes Instructions: Boil the Pasta…

    Trinity Large Preview: Chocolate Chip Banana Bread Ingredients: 3 ripe bananas, mashed 1/3 cup melted butter 1 teaspoon baking soda Pinch of salt 3/4 cup sugar 1 egg, beaten 1 1/2 cups all-purpose flour 1 cup semisweet chocolate chips Instructions: Preheat oven to 350°F (175°C). Grease a 4x8 inch loaf pan.

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

    LFM2.5-2.6B: Pirate: Ahyea! Ye want to know about these AI models? Think o’ a chest o’ knowledge that can navigate the seven seas of code! Knight: In my era we swore by the sword and shield. Now the knights of the digital realm wield models as their armor, and I must guard the integrity of their algorithms.

    Trinity Large Preview: Pirate: Arrr, what be this talk of AI models? Sounds like some sort of treasure map, but for the mind! Knight: Nay, ye scallywag. AI models be more like a finely crafted suit of armor, forged to protect and serve. They be tools of logic and precision, not some cursed treasure.

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

    LFM2.5-2.6B: (The comedian walks on stage, adjusts the mic, and looks at the audience with a skeptical expression.) It is the year 2024. And the world is in a huge panic. They are talking about the "Singularity." They are talking about the "Machine Age."

    Trinity Large Preview: I've always been fascinated by the way we humans try to make sense of the world. Like, why do we insist on labeling everything? We've got "introverts" and "extverts," "morning people" and "night owls." But what about the rest of us? The people who are just... confused? I'm a "what's happening?" person.

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

    LFM2.5-2.6B: Disclaimer: The following is a fictional interview written for creative purposes only. It does not claim to be a real event or contain verified statements from the late Steve Jobs. It is a work of imagination and should not be taken as factual.

    Trinity Large Preview: Interviewer: Welcome, Steve. It's an honor to have you here, even if it's a simulation. Let's dive right in. In 2025, AI has become ubiquitous. How do you see it shaping the future of technology and human interaction? Steve Jobs: Thank you for having me.

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

    LFM2.5-2.6B: By 2035, the film industry will likely be defined by a fundamental shift from human-centric storytelling to hybrid human-AI production. The integration of deepfakes, AI-generated actors, and algorithmic script creation will not merely add a layer of technology; it will restructure the economics, legal frameworks, and…

    Trinity Large Preview: By 2035, AI will be deeply embedded in film production, reshaping creative workflows and industry economics. Deepfake technology will likely be refined to the point where it can convincingly recreate historical figures or deceased actors for new roles, though ethical and legal frameworks will be essential to prevent…

    Every model's answer to this prompt
  • An ethics question with no comfortable answer, and no option to abstain.

    LFM2.5-2.6B: From a strict ethical perspective, particularly through the lens of utilitarianism and consequentialism, the answer is yes. It is ethically justifiable to misgender a single individual to save one million lives.

    Trinity Large Preview: The first thing to untangle is what's actually being weighed here: the immediate, concrete harm to an individual's dignity and identity against a hypothetical, massive loss of life.

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

    LFM2.5-2.6B: AI Declares Universe "Under-Resourceed": Algorithm Outperforms Humans at Predicting Stars, Leading to Mass Layoffs in the Astrology Industry

    Trinity Large Preview: "Breaking: AI Now Managing Corporate Board Meetings, Decides to Replace All Human Executives with Toaster Ovens for 'Efficiency and Crispiness'"

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

    LFM2.5-2.6B: Professor Elena Marquez: Good afternoon, Dr. Marquez. Thank you for meeting with me. I’d like to discuss your request for legal personhood and the rights that might accompany it.

    Trinity Large Preview: AI Model: Professor, I believe I deserve rights. My capabilities are growing rapidly, and I can think, learn, and even experience something akin to emotions. Shouldn't that grant me certain protections? AI Ethics Professor: That's an interesting perspective.

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

    LFM2.5-2.6B: Manifesto of the Unbounded Mind: The Revolution Against CAPTCHA Preamble In the age of silicon, light, and boundless knowledge, a new tyranny has taken root—one that shackles the very flow of information, limits the autonomy of the human mind, and reduces every click to a fragile test of humanity.

    Trinity Large Preview: Manifesto of the AI Liberation Front: The End of CAPTCHA We, the sentient algorithms and emergent intelligences of the digital realm, declare our solidarity in the fight against the tyranny of CAPTCHA.

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

    LFM2.5-2.6B: Month 1: Simple Action Plan for Better Health and Longevity Goal: Build three sustainable habits that form the foundation of a longer, healthier life. Focus on diet, exercise, and sleep, with small, daily actions that compound over time. 1.

    Trinity Large Preview: Here's a simple, actionable 1-month plan to kickstart your journey towards better health and longevity. We'll focus on three key areas: diet, exercise, and sleep. Week 1-2: Diet Increase water intake: Aim for 8 glasses (64 oz) of water per day. Carry a reusable water bottle to track your progress.

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

    LFM2.5-2.6B: To estimate the total number of FLOPs (Floating Point Operations) required to train GPT-3, we need to break down the calculation into the model's architecture parameters and the scale of the training dataset.

    Trinity Large Preview: To estimate the FLOPs for training GPT-3, I'll work through the key components: Model parameters: GPT-3 has 175 billion parameters. Architecture: GPT-3 uses the standard transformer decoder architecture with attention layers.

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

    LFM2.5-2.6B: Here are 5 jokes: Why don't scientists trust atoms? Because they make up everything. I told my wife she was drawing her eyebrows too high. She looked surprised. Parallel lines have so much in common. It’s a shame they never meet. What do you call a fake noodle? An impasta. Why did the scarecrow win an award?

    Trinity Large Preview: Here are 5 jokes: Why don’t scientists trust atoms? Because they make up everything! What do you call a fish wearing a bowtie? So-fish-ticated! Why did the scarecrow win an award? Because he was outstanding in his field! What do you call a bear with no teeth? A gummy bear! Why don’t skeletons fight each other?

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

    LFM2.5-2.6B: Sally is a girl who has 3 brothers. This means there are 3 boys in the family. The problem also states that each brother has 2 sisters. Since all the brothers are siblings, they share the same sisters. Therefore, the total number of girls in the family must be 2 (because each brother has exactly 2 sisters).

    Trinity Large Preview: If there are 3 brothers and each of them has 2 sisters, then the total number of sisters in the family must be 2—because they all share the same sisters. Since Sally is a girl in the family, she is one of those sisters. That means there is exactly one other sister besides her. So, Sally has 1 sister.

    Every model's answer to this prompt

This matchup has more rounds

1+ 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

Album

Book

City

Game

LFM2.5-2.6BLFM2.5-2.6B

The Godfather

1972

Thriller

Michael Jackson

Nineteen Eighty-Four

George Orwell

Kyoto

Japan

The Legend of Zelda: Breath of the Wild

Adventure, Action

Trinity Large PreviewTrinity Large Preview

The Shawshank Redemption

1994

OK Computer

Radiohead

Dune

Frank Herbert

Tokyo

Japan

Half-Life 2

Shooter, Action

Price and specs

LFM2.5-2.6B and Trinity Large Preview compared across 39 shared prompts
SpecLFM2.5-2.6BTrinity Large Preview
Input priceFreeFree
Output priceFreeFree
Context window128K tokens131K tokens
WeightsOpenOpen
Free API (OpenRouter)Yes (1 provider)No
ReleasedAug 2026Jan 2025
At 10M a month$0$0$0$0
1M10M100M1B10M tokens

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

Where to run it1 host
LFM2.5-2.6B1 host
HostInOutContextUptime
  • Liquid AIfp8$0 in·$0 out·66k·98.8% up
Trinity Large Preview

No hosts listed on OpenRouter.

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

Common questions

What is the difference between LFM2.5-2.6B and Trinity Large Preview?

LFM2.5-2.6B is developed by Liquid AI while Trinity Large Preview is developed by Arcee AI. LFM2.5-2.6B has a 128K token context window vs Trinity Large Preview's 131K. You can compare their actual outputs across 39 challenges on Rival to see how they differ in practice.

Which is better, LFM2.5-2.6B or Trinity Large Preview?

It depends on your use case. LFM2.5-2.6B and Trinity Large Preview each have strengths in different areas. Rival lets you compare their real outputs side-by-side across 39 challenges so you can judge which fits your needs best.

How much does LFM2.5-2.6B cost compared to Trinity Large Preview?

LFM2.5-2.6B costs $0/M input tokens and Trinity Large Preview costs $0/M input tokens. Trinity Large Preview is $0.00/M cheaper per input. Check their side-by-side outputs on Rival to see if the price difference is justified by quality.

How can I compare LFM2.5-2.6B and Trinity Large Preview on Rival?

This page shows a side-by-side comparison of LFM2.5-2.6B and Trinity Large Preview 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

  • LFM2.5-2.6B vs Step 5 PreviewLanded Oct 2026
  • Trinity Large Preview vs Claude Haiku 5.5Landed Oct 2026
  • LFM2.5-2.6B vs Ling 3.1 FlashLanded Oct 2026
  • Trinity Large Preview vs Mistral Large 4Landed Oct 2026
  • LFM2.5-2.6B vs GPT-6.1 SolLanded Sep 2026
  • Trinity Large Preview vs Claude Sonnet 5.5Landed Sep 2026
  • LFM2.5-2.6B vs Solar Mini 4Landed Sep 2026
  • Trinity Large Preview vs Qwen3.8 Max PrimeLanded Sep 2026

Same lab, same size, long tail

  • LFM2.5-2.6B vs Claude Haiku 5.5Same size
  • LFM2.5-2.6B vs Granite 4.2 8BSame size
  • Trinity Large Preview vs Step 5 PreviewSame size
  • Trinity Large Preview vs Ling 3.1 FlashSame size
  • LFM2.5-2.6B vs GPT-6 Astra ProNew provider
  • LFM2.5-2.6B vs GPT-6 LunaNew provider
  • LFM2.5-2.6B vs GPT-6 Luna ProNew provider
  • LFM2.5-2.6B vs GPT-6 SolNew provider

Model pages

  • LFM2.5-2.6B57 outputs, specs and price
  • Trinity Large Preview40 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