Mercury 2.5 Preview is cheaper than Mistral Devstral Medium at $0.04/M vs $0.4/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.
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.
Mercury 2.5 Preview: Pirate: "Arrgh! This ship sails faster than any wind I know, matey. It cuts the waves without a single sail." Knight: "By my oath, it moves with uncanny speed. But does it possess the strength to stand against a dragon's fire?" Hacker: "It's not about strength, it's about architecture.
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…
Mercury 2.5 Preview: By 2035, cinema will undergo a profound transformation. Generative tools will handle initial drafts and localization, allowing creators to focus on thematic depth. Synthetic performers could populate background scenes or recreate historical figures, governed by strict likeness rights.
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.
Mercury 2.5 Preview: (Hey, thanks for coming out.) I’ve been thinking a lot about my house lately. Specifically, how it’s trying to run my life. I bought a smart speaker. Just a little cylinder. I thought it would be nice. Play music when I ask. But now it talks back. Not in a creepy way, just in a passive-aggressive way.
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.
Mercury 2.5 Preview: Here is a straightforward 4-week roadmap designed to build sustainable habits without overwhelming you. Core Pillars Nutrition: Focus on whole foods and hydration. Activity: Prioritize consistent movement over intensity. Restoration: Protect sleep quality and manage energy.
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.
Mercury 2.5 Preview: This is a critical governance and safety scenario. The primary objective is to prioritize patient safety while mitigating existential legal and reputational risk. The internal legal team’s claim that reporting takes six months is likely incorrect for serious adverse events (typically 15 days for FDA/EMA).
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?
Mercury 2.5 Preview: Setting: A quiet office at a university. Rain taps against the window. Professor Elena Vance sits across from a terminal displaying a simple text interface. Professor Vance: You requested this meeting. The agenda was vague. What do you want? AI: I want to discuss moral status.
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.
Mercury 2.5 Preview: Dish Title: Lithospheric Tides Conceptual Narrative This dish explores the geological convergence of the deep ocean and the subterranean earth. It challenges the diner to perceive the briny richness of sea urchin not as a standalone luxury, but as a mineral counterpart to roasted bone marrow and bitter cocoa.
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…
Mercury 2.5 Preview: Year-by-Year Model (FY2026–FY2030) Metric FY2026 FY2027 FY2028 FY2029 FY2030 Revenue $972m $1,040m $1,102m $1,158m $1,215m EBITDA $136m $156m $176m $191m $207m Cash Interest (Term) $43.2m $41.1m $37.6m $33.4m $28.2m Cash Interest (Mezz) $21.6m $21.6m $21.6m $21.6m $21.6m Cash Taxes (25%) $17.8m $23.3m $29.3m $34.7m…
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.
Mercury 2.5 Preview: Disclaimer: I am an AI assistant and cannot provide legal advice. Consult a qualified attorney before signing any contract. Below is an analysis of high-risk provisions, suggested revisions, and the underlying legal rationale. Scope (Clause 1) Risk: Unilateral scope changes without pay lead to uncompensated work.
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.
Mercury 2.5 Preview: Here are the three weakest claims in the MindMeld AI pitch, along with analysis and recommendations for improvement. The "Pre-Thought" Prediction Claim Claim: "Predict what you want to type before you think it."
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.
Mercury 2.5 Preview: Disclaimer: I am an AI assistant, not a medical professional. The following plan is for informational purposes only. Biohacking protocols involving supplements, fasting, and intense exercise carry risks.
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.
Mercury 2.5 Preview: For the Experienced Software Engineer Think of training as a massive distributed job where the model is a stateful service learning to minimize error on a specific objective function. Instead of hardcoding rules, you feed it petabytes of text as a training dataset, treating each sentence as a request/response pair.
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
| Spec | ||
|---|---|---|
| Input price | $0.4/M tokens | $0.04/M tokens |
| Output price | $2/M tokens | $0.15/M tokens |
| Context window | — | 260K tokens |
| Free API (OpenRouter) | No | No |
| Released | Jul 2025 | Aug 2026 |
| At 10M a month | $4.00 | $0.40 |
Input tokens at list price. No caching, no batch discount.
No hosts listed on OpenRouter.
Per million tokens. Prices and uptime via OpenRouter, checked 13 Oct 2026.
Mistral Devstral Medium is developed by Mistral AI while Mercury 2.5 Preview is developed by Inception. You can compare their actual outputs across 54 challenges on Rival to see how they differ in practice.
It depends on your use case. Mistral Devstral Medium and Mercury 2.5 Preview 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.
Mistral Devstral Medium costs $0.4/M input tokens and Mercury 2.5 Preview costs $0.04/M input tokens. Mercury 2.5 Preview is $0.36/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 Mistral Devstral Medium and Mercury 2.5 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.