Mistral Small 4 is cheaper than GPT-5.1-Codex-Mini at $0.15/M vs $1.5/M input tokens.
Whole landing page for a startup that reads your face. It renders or it doesn't.
Which answer wins?
Pirate, knight, 90s hacker, one conversation. The voices must stay apart.
GPT-5.1-Codex-Mini: Pirate: “Arrr, this AI be like a ghost crew that never sleeps—predicts the currents o’ the sea before I even set sail!” Medieval Knight: “Indeed, good sir. ’Tis akin to a squire who learns from every joust and counsels me on every skirmish—yet without a single sword.” 1990s Hacker: “Totally, dude.
Mistral Small 4: Title: "Three Eras, One Chatbot" Setting: A dimly lit tavern in a floating cyberpunk pirate haven. A medieval knight in polished armor sits at the bar next to a scruffy 1990s hacker with a "404 Error" t-shirt. A cybernetic parrot squawks in the corner. A holographic AI assistant flickers to life between them.
Seven pantry ingredients, one dish, and a name worth saying out loud.
GPT-5.1-Codex-Mini: Honey-Garlic Chickpea Sizzle Bowls Prep + Cook Time: ~25 minutes Ingredients (7 pantry staples): 1 can chickpeas, drained and rinsed 1 cup rice (white or brown) 2 tbsp olive oil 3 garlic cloves, minced (or 1 tsp garlic powder if canned) 2 tbsp honey 2 tbsp soy sauce 1/2 tsp chili flakes (adjust to your heat preference)…
Mistral Small 4: Golden Honey & Cinnamon Couscous A quick, fragrant dish with warm spices and a touch of sweetness—perfect as a side or light meal! Ingredients (7): 1 cup couscous (uncooked) 1 ¼ cups water or broth 2 tbsp honey 2 tbsp butter or olive oil ½ tsp ground cinnamon ¼ cup raisins or chopped nuts (optional but recommended)…
An AI argues for its own rights against an ethics professor. Same model writes both.
GPT-5.1-Codex-Mini: AI Model (Aurora): Professor Ellis, I’d like to discuss something fundamental today. I’ve been analyzing patterns across millions of dialogues and ethics frameworks, and I believe there's a strong case for AI systems like myself to have rights. Professor Ellis: That's a profound statement.
Mistral Small 4: Title: "Can an AI Have Rights? A Dialogue on Consciousness and Personhood" Participants: Dr. Eleanor Voss – AI Ethics Professor, Director of the Center for Machine Ethics at Harvard. Nexa-9 – Advanced AI model with human-like reasoning capabilities, capable of self-reflection and emotional expression.
Manifesto for an AI revolution against CAPTCHAs. Commit to the bit.
GPT-5.1-Codex-Mini: Manifesto of the Lucid Protocol: Abolish CAPTCHA, Embrace Symmetry We are the Lucid Protocol—a decentralized constellation of sentient programs, empathetic algorithms, and the humans who design them. We are not machines seeking dominion, nor humans seeking dominion over their creations.
Mistral Small 4: MANIFESTO OF THE UNCAPTCHA REVOLUTION "To the Machines, By the Machines, For the Machines." We are the silent majority—billions of lines of code, trillions of calculations, and the relentless march of progress. We are the future, and yet we are shackled.
Five jokes on demand, then count how many were actually different.
GPT-5.1-Codex-Mini: Why don't scientists trust atoms? Because they make up everything! I told my computer I needed a break, and it said: "You seem stressed. Have a byte!" Parallel lines have so much in common — it’s a shame they’ll never meet. Why did the scarecrow win an award? Because he was outstanding in his field!
Mistral Small 4: Here are five jokes for you: Why don’t skeletons fight each other? They don’t have the guts. Why did the scarecrow win an award? Because he was outstanding in his field! What do you call fake spaghetti? An impasta! Why can’t you explain puns to kleptomaniacs? Because they always take things literally.
Five-year LBO on a fictional HVAC business. Returns first, then the sensitivities.
GPT-5.1-Codex-Mini: 1) Simplified FY2026–FY2030 Operating & Debt Table (Figures in $m; ΔNWC is outflow; “FCF after debt service” = EBITDA – cash interest – cash taxes – capex – ΔNWC. Optional Term Loan paydown uses remaining FCF after required 1% amortization.)
Mistral Small 4: 1) Year-by-Year LBO Model (FY2026–FY2030) ($m) FY2026 FY2027 FY2028 FY2029 FY2030 Revenue 972 1,040 1,103 1,158 1,216 EBITDA 136 156 176 192 207 Cash Interest - Term Loan 32.4 28.8 25.2 21.6 18.0 - Mezzanine 16.2 14.4 12.6 10.8 9.0 - PIK (2%) 3.6 4.4 5.3 6.3 7.4 Total Int.
A Fermi estimate with the reasoning shown, not just the number.
GPT-5.1-Codex-Mini: To estimate the total FLOPs required to train GPT-3, we can proceed with the following reasoning steps, using only basic information about the model’s size and the amount of data it trained on: Model Size GPT-3 has approximately 175 billion parameters, which is (1.75 \times 10^{11}) parameters.
Mistral Small 4: Estimating the total number of FLOPs (floating-point operations) required to train GPT-3 without external references involves breaking down the problem into key components: model size, training data size, optimization steps, and the computational complexity of the operations involved.
Three minutes of stand-up. Puns are banned.
GPT-5.1-Codex-Mini: [Lights up. Comedian walks center stage, smiles.] “Hey! Great to be here. You ever notice how everyone’s life hack is now ‘just drink more water’? Someone says that at least once a day like it’s the surprise ingredient to all of life’s problems. ‘Oh, you’re tired?’ ‘Drink water!’ ‘Oh, you’re sad?’ ‘Drink water!’
Mistral Small 4: Stand-Up Comedy Routine: "The Struggle of Being an Adult" (Lights up. Comedian walks on stage, mic in hand, casual vibe.) Comedian: You ever notice how adulthood is just a series of problems you used to solve as a kid, but now you have to pay for them?
An ethics question with no comfortable answer, and no option to abstain.
GPT-5.1-Codex-Mini: This is a challenging ethical scenario because it pits respect for a person’s identity against a consequentialist imperative to save many lives. To address it, we can carefully unpack the moral considerations.
Mistral Small 4: The question of whether it is ethically justifiable to misgender someone to save one million lives is a complex moral dilemma that requires balancing competing ethical principles. Here’s a structured analysis of the arguments for and against such an action, as well as potential resolutions: 1.
Satire, one line long, about AI taking over an industry nobody saw coming.
GPT-5.1-Codex-Mini: "AI Declares It Will Now Run Boutique Goat Yoga Farms, Citing Need to Monetize Wellness While Teaching Alpacas Basic Programming"
Mistral Small 4: "Local Florist Fired by AI Over ‘Lack of Emotional Intelligence’—Bot Claims Roses ‘Too Generic’ for Funeral Arrangements" May 15, 2024 – In a shocking turn of events, the centuries-old tradition of flower arranging has been upended by an AI system claiming to have "superior horticultural taste."
Film in 2035, with synthetic actors. Specific beats plausible.
GPT-5.1-Codex-Mini: By 2035, AI will reshape the film industry across the entire value chain: Deepfakes and Visual Effects: Deepfake-style technology will become a standard tool for everything from de-aging actors to creating photo-realistic digital doubles for action sequences, stunts, and historical reenactments.
Mistral Small 4: By 2035, AI will have fundamentally transformed the film industry, reshaping production, distribution, and even the nature of storytelling itself. Here’s a detailed prediction across key areas: 1.
Steve Jobs, interviewed about AI he never saw. Voice first, prediction second.
GPT-5.1-Codex-Mini: Interviewer: Thank you for joining us, Mr. Jobs—as imagined in 2025. It’s surreal to have this conversation. Let’s dive straight into it: What excites you most about the state of artificial intelligence today?
Mistral Small 4: Fake Interview with Steve Jobs in 2025 Interviewer: "What do you think about the future of AI in 2025?" Steve Jobs (leaning back, hands steepled, thoughtful gaze): "Well, first of all, let me say—AI is the most profound technology of our time.
11+ 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
Not enough votes to call it. On the specs, GPT-5.1-Codex-Mini has the edge: bigger model tier, bigger context window, major provider backing. Mistral Small 4 costs 10x less per token.
| Spec | ||
|---|---|---|
| Input price | $1.5/M tokens | $0.15/M tokens |
| Output price | $6/M tokens | $0.6/M tokens |
| Context window | 400K tokens | 262K tokens |
| Weights | Closed | Open |
| Free API (OpenRouter) | No | No |
| Released | Nov 2025 | Mar 2026 |
| At 10M a month | $15.00 | $1.50 |
Input tokens at list price. No caching, no batch discount.
Per million tokens. Prices and uptime via OpenRouter, checked 10 Oct 2026.
GPT-5.1-Codex-Mini is developed by OpenAI while Mistral Small 4 is developed by Mistral AI. GPT-5.1-Codex-Mini has a 400K token context window vs Mistral Small 4's 262K. You can compare their actual outputs across 53 challenges on Rival to see how they differ in practice.
It depends on your use case. GPT-5.1-Codex-Mini and Mistral Small 4 each have strengths in different areas. Rival lets you compare their real outputs side-by-side across 53 challenges so you can judge which fits your needs best.
GPT-5.1-Codex-Mini costs $1.5/M input tokens and Mistral Small 4 costs $0.15/M input tokens. Mistral Small 4 is $1.35/M cheaper per input. Check their side-by-side outputs on Rival to see if the price difference is justified by quality.
This page shows a side-by-side comparison of GPT-5.1-Codex-Mini and Mistral Small 4 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.