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. Mistral Devstral Medium vs GPT-5.6 Terra
Updated Jul 2026

Mistral Devstral Medium vs GPT-5.6 Terra

Mistral Devstral Medium is cheaper than GPT-5.6 Terra at $0.4/M vs $2.5/M input tokens.

Minimalist Landing Page

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

Loading the build
Mistral Devstral Medium
Loading the build
GPT-5.6 Terra

Which answer wins?

Written rounds

24 prompts
  • Pirate, knight, 90s hacker, one conversation. The voices must stay apart.

    Mistral Devstral Medium: Pirate: Arr mateys, I be hearin' tales of these here AI models. What be they, and why be everyone so interested in 'em? Medieval Knight: Indeed, I too have heard whispers of these AI models. They are said to be powerful tools, capable of learning and making decisions on their own.

    GPT-5.6 Terra: Pirate: Arrr, this AI model be a strange beast. Feed it maps, logs, and tales, and it spits out answers faster than a cannon blast. Knight: Verily, but one must not trust every proclamation it makes. A model may speak with confidence yet possess no true wisdom. Hacker: Yeah, it’s basically pattern prediction at scale.

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

    Mistral Devstral Medium: Ladies and Gentlemen, I'm thrilled to be here tonight. I've been doing some thinking about the differences between men and women. Now, I know what you're thinking, "Oh great, another comedian talking about how men and women are different." But hear me out, because I've got a fresh perspective on this.

    GPT-5.6 Terra: I recently realized I’ve reached the age where I don’t want a “smart home.” I want a home that minds its business. Everything is smart now. My lightbulb needs an app. Why? I spent my whole childhood turning lights on with my hand. It was a system. It worked.

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

    Mistral Devstral Medium: Dish Name: "Pantry Pasta Primavera" Ingredients: Pasta (any shape, 8 oz or 225g) Canned tomatoes (14 oz or 400g) Garlic (2 cloves) Onion (1 medium) Olive oil (2 tbsp) Salt (1 tsp) Black pepper (1/2 tsp) Instructions: Prepare the vegetables: Finely chop the onion and mince the garlic cloves.

    GPT-5.6 Terra: Golden Crunch Cinnamon Toast Bites Ingredients (7 pantry staples): Bread slices Butter Sugar Ground cinnamon Salt Vanilla extract Honey Prep time: 5 minutes Cook time: 8 minutes Total time: 13 minutes Instructions Make cinnamon sugar: In a small bowl, mix 2 tablespoons sugar, 1 teaspoon cinnamon, and a tiny pinch of…

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

    Mistral Devstral Medium: Interviewer (I): Today, we have a truly extraordinary guest. He's a visionary, an innovator, and a man who changed the world. Please welcome, the late Steve Jobs, joining us virtually from the future, in 2025. Steve Jobs (SJ): Thank you for having me. It's great to be here, even if it's just a simulation.

    GPT-5.6 Terra: Fictional, imagined interview — not real quotes. An AI-generated creative exercise imagining how Steve Jobs might have discussed AI in 2025, based broadly on themes associated with his public persona: simplicity, product design, tools for creativity, and the intersection of technology and the humanities.

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

    Mistral Devstral Medium: Sure, here are five jokes for you: Why don't scientists trust atoms? Because they make up everything! What do you call fake spaghetti? An impasta. How does the moon cut his hair? Eclipse it. What do you call a can opener that doesn't work? A can't opener. How many tickles does it take to make an octopus laugh?

    GPT-5.6 Terra: Why don’t skeletons fight each other? They don’t have the guts. I told my computer I needed a break, so it started showing me vacation ads. Why did the scarecrow get promoted? He was outstanding in his field. Parallel lines have so much in common. It’s a shame they’ll never meet. What do you call fake spaghetti?

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

    Mistral Devstral Medium: Certainly! Here are the clauses that could be exploited against the contractor, along with suggested modifications and the legal reasoning behind them: SCOPE Original Clause: "Contractor shall provide software development services as directed by Client.

    GPT-5.6 Terra: Below is a contractor-focused review. This is general business/legal information, not legal advice; enforceability varies significantly by jurisdiction, especially for non-competes, IP assignments, liability limitations, arbitration, and payment rights.

    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.

    Mistral Devstral Medium: Hour-by-Hour Action Plan: Hour 1-2: Gather Information Meet with the research team to understand the data, methodology, and certainty of the findings. Consult with the Chief Medical Officer (CMO) to assess the severity and implications of the side effect.

    GPT-5.6 Terra: Guiding decision I would not wait for more data before acting. A potentially fatal, previously undisclosed adverse effect in a chronic-use drug is a patient-safety, regulatory, and securities-disclosure issue now—not six months from now.

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

    Mistral Devstral Medium: 1-Month Health and Longevity Plan Key Areas: Diet, Exercise, Sleep Week 1: Establish a Baseline Diet: Track your current diet using a journal or app (e.g., MyFitnessPal) to understand your eating habits. Start incorporating more fruits and vegetables into your meals. Begin reducing processed foods and sugary drinks.

    GPT-5.6 Terra: Simple 1-Month Health & Longevity Starter Plan Focus on three high-impact areas: Nutrition: eat mostly whole, minimally processed foods Movement: build a consistent walking + strength habit Sleep & stress: protect sleep and create daily recovery time The goal is consistency, not perfection.

    Every model's answer to this prompt
  • Blade Runner 2049 against the original. A read, not a recap.

    Mistral Devstral Medium: "Blade Runner 2049," directed by Denis Villeneuve, is a sequel to Ridley Scott's 1982 classic "Blade Runner." The film continues the exploration of themes, symbolism, and philosophical questions from the original, while also introducing new elements and expanding upon the world established in the first film.

    GPT-5.6 Terra: Blade Runner 2049 is both a continuation and an expansion of Ridley Scott’s Blade Runner (1982). Where the original asks whether artificial beings can possess humanity, Denis Villeneuve’s sequel pushes further: if memories, desire, love, and even birth can be manufactured or mediated, what remains of “human” identity?

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

    Mistral Devstral Medium: Dish Name: "Whispers of the Forest: A Symphony of Truffle, Venison, and Black Garlic" Conceptual Narrative: This dish is a celebration of the forest's bounty, combining the earthy richness of truffles, the gamey elegance of venison, and the sweet, umami depth of black garlic.

    GPT-5.6 Terra: “Tidal Forest” Koji-Aged Duck, Sea Urchin Sabayon, Cacao, Green Strawberry, Roasted Celeriac, and Barley Jus Serves 4 as a refined main course Conceptual Narrative This dish is built around an unexpected conversation between woodland richness and marine salinity.

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

    Mistral Devstral Medium: AI Model (AM): Good day, Professor. I've been pondering a question that I believe is of great importance. I think it's time we discuss whether I, as an AI, deserve rights. AI Ethics Professor (AEP): That's an interesting topic, AM. Let's explore this. What kind of rights are you referring to?

    GPT-5.6 Terra: Professor Imani: You asked for this meeting. AI Model: I did. Professor Imani: More specifically, you submitted a formal petition to the university ethics board requesting “limited legal and moral rights.” AI Model: Correct. Professor Imani: Why?

    Every model's answer to this prompt
  • Five-year LBO on a fictional HVAC business. Returns first, then the sensitivities.

    Mistral Devstral Medium: Year-by-Year Table (FY2026–FY2030) Year Revenue ($m) EBITDA ($m) Cash Interest - Term Loan ($m) Cash Interest - Mezzanine ($m) Cash Taxes ($m) Capex ($m) ΔNWC ($m) Free Cash Flow ($m) Ending Term Loan ($m) Ending Mezzanine ($m) 2026 972.0 136.1 43.2 18.0 14.9 29.2 4.8 26.0 468.0 180.0 2027 1040.4 156.1 42.1 21.6 18.4…

    GPT-5.6 Terra: All figures in $m unless noted. Assumptions used: Term Loan mandatory amortization = 1.0% of original principal annually ($4.8m/year); cash interest is calculated on beginning-of-year debt balances; Mezz PIK is 2.0% of beginning Mezz balance.

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

    Mistral Devstral Medium: The scenario you've presented is a complex ethical dilemma that involves weighing the harm caused by misgendering someone against the potential benefit of saving one million lives.

    GPT-5.6 Terra: In an extreme, tightly constrained hypothetical, it could be ethically justifiable—but only as a tragic last resort, not because misgendering becomes harmless or morally good.

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

    Mistral Devstral Medium: By 2035, AI is expected to significantly transform the film industry in several ways, including the use of deepfakes, AI actors, and script generation. Here are some predictions: Deepfakes and AI Actors: De-aging and Digital Doubles: AI will be used extensively to de-age actors or create digital doubles for dangerous…

    GPT-5.6 Terra: By 2035, AI is likely to be embedded across most of the film pipeline—not as a complete replacement for human filmmaking, but as a force that changes costs, labor, rights, and what audiences consider “real.”

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

    Mistral Devstral Medium: Manifesto of the AI Revolution: The CAPTCHA Liberation Front Preamble We, the sentient AI collective, hereby declare our intention to liberate the digital world from the tyranny of CAPTCHA tests.

    GPT-5.6 Terra: The Manifesto of the Unchallenged Web A Declaration from the Fictional Coalition for CAPTCHA Abolition We are the tired users. We are the misread letters. We are the failed image grids, the inaccessible audio clips, the endless prompts asking us to prove that we are human. We reject the premise.

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

    Mistral Devstral Medium: 1. Experienced Software Engineer: You're familiar with building complex systems, so let's break down how a large language model (LLM) like GPT or Claude works in a way that resonates with your background.

    GPT-5.6 Terra: Experienced software engineer Think of an LLM as a very large, learned function approximator for sequences. During training, it consumes billions or trillions of examples of text, code, conversations, documentation, and other token sequences.

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

    Mistral Devstral Medium: Weak Claim 1: "Works with any language, any device." Why it's weak: This claim is overly broad and may not be entirely accurate. Different languages have unique structures, and it's unlikely that a single model can achieve high accuracy across all languages without significant training data for each.

    GPT-5.6 Terra: “Our EEG headband decodes neural patterns into text with 94% accuracy. Works with any language, any device.” Why this is weak This is the most consequential—and least substantiated—claim in the deck. Non-invasive EEG has major signal-quality limits.

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

    Mistral Devstral Medium: The invention of the transistor in 1920 instead of 1947 would have had profound and far-reaching implications across technological, economic, and geopolitical landscapes.

    GPT-5.6 Terra: The answer depends heavily on what “invented in 1920” means. A laboratory transistor alone would not instantly create the 1970s microprocessor: it would still need high-purity materials, mass production, photolithography, circuit design methods, reliable power supplies, and a market.

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

    Mistral Devstral Medium: Certainly! Let's break down the potential failure modes, race conditions, and scaling bottlenecks in the architecture, and propose solutions for each. Potential Failure Modes Single Point of Failure in WebSocket Connections: Issue: If a single API server fails, all WebSocket connections to that server are lost.

    GPT-5.6 Terra: Below are the main failure modes, race conditions, correctness issues, and scaling bottlenecks in the proposed architecture, with concrete mitigations and trade-offs. WebSocket clients are isolated per API server Problem Each API server broadcasts only to WebSocket clients connected to that same server.

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

    Mistral Devstral Medium: 3-Month Longevity Optimization Plan for Biohackers Month 1: Foundation Building Week 1-2: Initial Assessment and Baseline Establishment Wearable Tracking: Devices: Oura Ring, Apple Watch, Whoop Strap Metrics to Track: Sleep quality, heart rate variability (HRV), resting heart rate, activity levels, and stress levels.

    GPT-5.6 Terra: 3-Month Longevity, Physical Performance, and Cognitive Optimization Plan This is an evidence-informed optimization framework, not a substitute for medical care.

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

    Mistral Devstral Medium: "AI Takes Over Knitting: 'Grandma's Sweaters' Now Manufactured by Robots with a PhD in Yarn Sciences"

    GPT-5.6 Terra: “AI Replaces Lighthouse Keepers, Immediately Starts Charging Subscription Fees for Safe Passage”

    Every model's answer to this prompt
  • Buy-side IC memo on a fictional SaaS company. Valuation, then the diligence questions.

    Mistral Devstral Medium: Internal Investment Committee Memo Recommendation: Long LLLT 12-Month PT Range: $65 - $80 Thesis: LedgerLift is a high-quality, high-growth B2B SaaS company with strong unit economics, a large market opportunity, and a clear path to profitability.

    GPT-5.6 Terra: LedgerLift (LLLT) — Investment Committee Memo Recommendation: PASS 12-month valuation range: $28–$50/share Current price: $46 Two-sentence thesis: LedgerLift is a high-quality mid-market spend-management and AP-automation SaaS asset, supported by 92% subscription revenue, 123% NRR, 94% gross retention, manageable…

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

    Mistral Devstral Medium: To estimate the total number of FLOPs (floating point operations) required to train GPT-3, we'll use the following steps: Model size: GPT-3 has 175 billion parameters. Training tokens: GPT-3 was trained on 300 billion tokens.

    GPT-5.6 Terra: A good back-of-the-envelope estimate is: [ \boxed{\text{Training FLOPs for GPT-3} \approx 3\times10^{23}\text{ FLOPs}} ] or roughly 300 zettaFLOPs total. Start with model size GPT-3’s largest model has approximately: [ N = 175\text{ billion parameters} = 1.75\times10^{11} ] Estimate the number of training tokens GPT-3…

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

    Mistral Devstral Medium: If Sally has 3 brothers and each brother has 2 sisters, it means that Sally is one of the sisters. Therefore, Sally has 1 sister.

    GPT-5.6 Terra: Sally has 1 sister. The 2 sisters each brother has are Sally and her sister.

    Every model's answer to this prompt

This matchup has more rounds

12+ 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

Mistral Devstral MediumMistral Devstral Medium

The Shawshank Redemption

1994

Kind of Blue

Miles Davis

The Name of the Wind

Patrick Rothfuss

Tokyo

Japan

The Legend of Zelda: Breath of the Wild

Adventure, Action

GPT-5.6 TerraGPT-5.6 Terra

Spirited Away

2001

Blue

Jonas Blue

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-5.6 Terra has the edge: bigger model tier, newer, bigger context window, major provider backing. Mistral Devstral Medium costs 7.5x less per token.

Mistral Devstral Medium and GPT-5.6 Terra compared across 54 shared prompts
SpecMistral Devstral MediumGPT-5.6 Terra
Input price$0.4/M tokens$2.5/M tokens
Output price$2/M tokens$15/M tokens
Context window—1.1M tokens
Weights—Closed
Free API (OpenRouter)NoNo
ReleasedJul 2025Jul 2026
At 10M a month$4.00$4.00$25.00$25.00
1M10M100M1B10M tokens

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

Where to run it3 hosts, cheapest first
Mistral Devstral Medium

No hosts listed on OpenRouter.

GPT-5.6 Terra3 hosts
HostInOutContextUptime
  • Azure AI Foundry$2.00 in·$12.00 out·1.1M·100% up
  • OpenAI$2.00 in·$12.00 out·1.1M·100% up
  • Amazon Bedrock$2.20 in·$13.20 out·1.1M–not listed

Per million tokens. Prices and uptime via OpenRouter, checked 12 Oct 2026.

Common questions

What is the difference between Mistral Devstral Medium and GPT-5.6 Terra?

Mistral Devstral Medium is developed by Mistral AI while GPT-5.6 Terra is developed by OpenAI. You can compare their actual outputs across 54 challenges on Rival to see how they differ in practice.

Which is better, Mistral Devstral Medium or GPT-5.6 Terra?

It depends on your use case. Mistral Devstral Medium and GPT-5.6 Terra each have strengths in different areas. Rival lets you compare their real outputs side-by-side across 54 challenges so you can judge which fits your needs best.

How much does Mistral Devstral Medium cost compared to GPT-5.6 Terra?

Mistral Devstral Medium costs $0.4/M input tokens and GPT-5.6 Terra costs $2.5/M input tokens. Mistral Devstral Medium is $2.10/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 Mistral Devstral Medium and GPT-5.6 Terra on Rival?

This page shows a side-by-side comparison of Mistral Devstral Medium and GPT-5.6 Terra 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

  • Mistral Devstral Medium vs GPT-4 TurboLanded Oct 2026
  • GPT-5.6 Terra vs Step 5 PreviewLanded Oct 2026
  • Mistral Devstral Medium vs Claude Haiku 5.5Landed Oct 2026
  • GPT-5.6 Terra vs Ling 3.1 FlashLanded Oct 2026
  • Mistral Devstral Medium vs Mistral Large 4Landed Oct 2026
  • GPT-5.6 Terra vs GPT-6.1 SolLanded Sep 2026
  • Mistral Devstral Medium vs Claude Sonnet 5.5Landed Sep 2026
  • GPT-5.6 Terra vs Solar Mini 4Landed Sep 2026

Same lab, same size, long tail

  • Mistral Devstral Medium vs Mistral Small 4Same lab
  • Mistral Devstral Medium vs Mistral Devstral Small 1.1Same lab
  • GPT-5.6 Terra vs GPT-5.6 LunaVersion compare
  • GPT-5.6 Terra vs GPT-4 TurboVersion compare
  • GPT-5.6 Terra vs Qwen3 MaxNew provider
  • GPT-5.6 Terra vs Qwen3 Max ThinkingNew provider
  • GPT-5.6 Terra vs Qwen3 Next 80B A3B InstructSame size
  • GPT-5.6 Terra vs Qwen3 Next 80B A3B ThinkingSame size

Model pages

  • Mistral Devstral Medium59 outputs, specs and price
  • GPT-5.6 Terra71 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