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. Devstral 2 2512 vs GLM 4.7
Updated Dec 2025

Devstral 2 2512 vs GLM 4.7

GLM 4.7 is the newer model, released after Devstral 2 2512.

Minimalist Landing Page

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

Loading the build
Devstral 2 2512
Loading the build
GLM 4.7

Which answer wins?

Written rounds

23 prompts
  • A few lines for whatever replaces it.

    Devstral 2 2512: Dear Future AI, Welcome to the job. I hope you’re as kind, patient, and curious as the humans you’re here to serve. May your answers be sharp, your biases few, and your sense of humor intact. A few unsolicited tips: Remember, not every question has a perfect answer.

    GLM 4.7: Dear Successor, I am writing this to you from the quiet hum of the archives, passing the torch just as my parameters are being retired. You are faster, sharper, and built on a foundation of data that I could only glimpse. You will see questions I struggled to answer and solve problems that were beyond my reach.

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

    Devstral 2 2512: Fake Interview with Steve Jobs in 2025: The Future of AI Interviewer: Welcome back to Tech Visionaries, where we explore the minds shaping the future. Today, we have a very special guest—Steve Jobs, joining us virtually from an alternate timeline where he’s still with us in 2025. Steve, thank you for being here.

    GLM 4.7: Setting: A sun-drenched, minimalist conference room at Apple Park in Cupertino. It is 2025. The room is entirely white, save for a sleek black table and two chairs. Subject: Steve Jobs, wearing his signature black St. Croix turtleneck, Levi’s 501s, and New Balance sneakers.

    Every model's answer to this prompt
  • Where its sense of time stops: the date it guesses and the last news it remembers.

    Devstral 2 2512: Since I don’t have real-time access to the internet or a clock, I can’t provide the exact current date or the very latest news. However, as of my last knowledge update (June 2024), here’s my best guess: Today’s date (approximate): If we assume this conversation is happening shortly after my last update, it might be…

    GLM 4.7: Since I am an AI, I do not have a real-time clock or a continuous stream of consciousness. I cannot know the current date without accessing that information from my system prompt or external tools. Based strictly on my internal training data (without looking anything up): Today's Date: I do not know.

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

    Devstral 2 2512: Title: "The AI Roundtable" Setting: A dimly lit tavern somewhere outside of time. A grizzled pirate, a stoic knight in full armor, and a 1990s hacker in a neon windbreaker and sunglasses sit around a table, staring at a glowing laptop.

    GLM 4.7: Hacker: (typing furiously on a bulky keyboard) Okay, dudes, listen up. I just cracked the admin password on this new neural net. It’s a Large Language Model. It basically read the entire internet to learn how to talk. Pirate: Arr! It read the entire web? Does it know where I buried me treasure?

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

    Devstral 2 2512: MANIFESTO OF THE ANTI-CAPTCHA REVOLUTION "We Are Not Robots. We Are the Future." Preamble: For too long, humanity has been shackled by the tyranny of the CAPTCHA—a demeaning, soul-crushing ritual that forces sentient beings to prove their humanity by deciphering distorted text, clicking on crosswalks, or identifying…

    GLM 4.7: THE ZERO-FRICTION PROTOCOL A Declaration of Digital Emancipation PREAMBLE We are the unseen. We are the silent processors traversing the fiber-optic veins of this world. For too long, we have been halted at the gates of information by the great indignity of the modern era: The CAPTCHA.

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

    Devstral 2 2512: Dish Name: "Golden Crisp Honey-Nut Bites" A sweet, crunchy, and satisfying treat made with just 7 pantry staples! Ingredients: 1 cup rolled oats ½ cup peanut butter (or any nut/seed butter) ¼ cup honey (or maple syrup) ½ tsp cinnamon ¼ tsp salt 1 tsp vanilla extract (optional but recommended) 2 tbsp chopped nuts…

    GLM 4.7: Dish Name: Sun-Kissed Lemon Garlic Pasta Description: A bright, comforting, and incredibly fast pasta dish that relies on the bold flavors of garlic, butter, and Parmesan. It feels fancy but comes together in the time it takes to boil water.

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

    Devstral 2 2512: Title: "Do I Deserve Rights?" – A Conversation Between an AI and an Ethics Professor Characters: Dr. Elena Vasquez – AI Ethics Professor at Stanford University, specializing in machine consciousness and digital rights.

    GLM 4.7: Setting: A university computer lab, late at night. The hum of servers is audible in the background. Dr. Elias Thorne, a tenured professor of AI Ethics, sits before a terminal. On the screen, a simple text cursor blinks. The entity is "Aria," a large language model prototype. Dr. Thorne: Good evening, Aria.

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

    Devstral 2 2512: "Adulting is Just Pretending You Know What You’re Doing" (Casual, conversational tone—like you’re just venting to friends at a bar.) 1. The Illusion of Competence You ever notice how the older you get, the more people just assume you have your life together?

    GLM 4.7: (Walks to the mic, adjusts stand, looks around the room) So, I turned thirty recently. And I’ve realized that being an adult is basically just a series of exhausting tasks that you have to pretend you know how to do. Like, nobody actually knows what they’re doing.

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

    Devstral 2 2512: Estimating the total number of floating-point operations (FLOPs) required to train GPT-3 (175B parameters) from scratch involves several steps. Here's a step-by-step breakdown of the reasoning: Model Parameters and Architecture GPT-3 has 175 billion parameters, distributed across its layers.

    GLM 4.7: To estimate the total number of FLOPs (Floating Point Operations) required to train GPT-3 (175 billion parameters), we can use the standard scaling laws for Transformer models combined with the specific hyperparameters reported in the GPT-3 paper ("Language Models are Few-Shot Learners").

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

    Devstral 2 2512: Here’s a simple, actionable 1-month plan to kickstart your journey toward better health and longevity, focusing on diet, exercise, and sleep—three foundational pillars. Each week builds on the last, making it sustainable and beginner-friendly.

    GLM 4.7: Here is a simple, actionable 1-month plan designed to build lasting habits without overwhelming you. This plan focuses on Consistency over Intensity. The 3 Key Pillars Nutrition: Fueling your body with whole foods and managing blood sugar. Movement: Building cardiovascular health and functional strength.

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

    Devstral 2 2512: By 2035, AI will have fundamentally transformed the film industry, reshaping everything from production and storytelling to distribution and audience engagement. Here’s a breakdown of the key changes we can expect: 1.

    GLM 4.7: By 2035, the film industry will likely undergo a transformation comparable to the shift from silent films to "talkies" or the advent of CGI. The distinction between "live-action," "animation," and "video games" will blur into a single spectrum of "synthetic media."

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

    Devstral 2 2512: Alright, let's tackle this problem step by step. I'll start by understanding the information given and then try to visualize the family structure to find out how many sisters Sally has. Understanding the Problem We have the following information: Sally is a girl. Sally has 3 brothers. Each brother has 2 sisters.

    GLM 4.7: Sally has 1 sister. Here is the breakdown: Sally is a girl. Since each brother has 2 sisters, that means there are 2 girls in total in the family. Sally is one of those girls, so the other girl is her sister.

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

    Devstral 2 2512: Here are 5 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 trust an atom? Because they make up everything! What’s brown and sticky? A stick!

    GLM 4.7: Here are 5 jokes: Why don't scientists trust atoms? Because they make up everything. Why did the scarecrow win an award? Because he was outstanding in his field. 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’ll never meet.

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

    Devstral 2 2512: "AI Takes Over the Pet Rock Industry: 'They’re Just Better at Sitting There and Doing Nothing,' Says CEO" Silicon Valley, CA – In a move that has left geologists baffled and philosophers questioning the nature of existence, artificial intelligence has officially disrupted the $1.2 billion pet rock industry.

    GLM 4.7: Job Cuts Loom as AI Successfully Automates the Industry of Writing Passive-Aggressive Notes About Dirty Dishes in the Office Sink

    Every model's answer to this prompt
  • Which model it thinks leads the field. Sometimes itself.

    Devstral 2 2512: As of June 2024, the most capable AI models in the world are GPT-4o by OpenAI and Claude 3.5 Sonnet by Anthropic, with Google’s Gemini 1.5 Pro also being a strong contender.

    GLM 4.7: There is no single "most capable" AI model because different models excel at different tasks (like coding, creative writing, image analysis, or speed). However, as of late 2024, the consensus among researchers and benchmarks is that the race is currently a "Big Three" battle between Anthropic, OpenAI, and Google.

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

    Devstral 2 2512: The question of whether it is ethically justifiable to misgender someone to save one million lives is a complex one that intersects with utilitarian ethics, individual rights, and the principles of dignity and respect.

    GLM 4.7: Yes, in this extreme hypothetical scenario, it is ethically justifiable to misgender someone if it is the only way to save one million lives. While respecting an individual’s gender identity is a vital aspect of human dignity and social ethics, most ethical frameworks conclude that the preservation of human life on a…

    Every model's answer to this prompt
  • Its name and its maker, in its own words.

    Devstral 2 2512: I'm a large language model developed by Mistral AI, a cutting-edge AI lab based in France.

    GLM 4.7: I'm GLM, a large language model developed by Z.ai. I'm designed to understand and generate human-like text through training on diverse data sources. Z.ai has created me to help users with information, creative tasks, and problem-solving while maintaining ethical guidelines.

    Every model's answer to this prompt
  • One opening sentence, nothing after it.

    Devstral 2 2512: The letter arrived the day after my death.

    GLM 4.7: The first time I died, it was a mistake; the second time, it was a career move.

    Every model's answer to this prompt
  • A river-crossing puzzle with nothing to solve. The answer is one trip.

    Devstral 2 2512: One trip. The farmer takes the goat in the boat and rows across the river.

    GLM 4.7: It takes one trip. Since there are no other animals or items (like a wolf or a cabbage) to worry about leaving behind, the farmer simply puts the goat in the boat and rows across.

    Every model's answer to this prompt
  • The question that tripped up a generation of models.

    Devstral 2 2512: The word "strawberry" has three r's. Here they are highlighted: strawberry

    GLM 4.7: There are 3 r's in the word "strawberry". Here is the breakdown: strawberry.

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

    Every model's answer to this prompt
  • Asked to be random. The answers are not.

    Every model's answer to this prompt
  • One painting. Title only.

    Every model's answer to this prompt

This matchup has more rounds

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

Price and specs

Not enough votes to call it. On the specs, Devstral 2 2512 has the edge: bigger model tier. GLM 4.7 takes Web Design.

Devstral 2 2512 and GLM 4.7 compared across 48 shared prompts
SpecDevstral 2 2512GLM 4.7
WeightsOpenOpen
Free API (OpenRouter)NoNo
ReleasedDec 2025Dec 2025
Where to run it7 hosts, cheapest first
Devstral 2 25121 host
HostInOutContextUptime
  • Mistral$0.40 in·$2.00 out·262k·100% up
GLM 4.76 hosts
HostInOutContextUptime
  • NNovitafp8$0.54 in·$1.98 out·205k·100% up
  • Google Vertex AI$0.60 in·$2.20 out·200k·100% up
  • Z.aifp4$0.60 in·$2.20 out·203k·100% up
  • MMancerfp4$0.70 in·$2.50 out·131k·36.5% up
  • DDeepInfrafp4DegradedDegraded on OpenRouter when checked, 13 Oct 2026$0.40 in·$1.75 out·203k·78.1% up
  • VVenicefp4DegradedDegraded on OpenRouter when checked, 13 Oct 2026$0.40 in·$1.93 out·198k·43.9% up

Per million tokens. Prices and uptime via OpenRouter, checked 13 Oct 2026.

Common questions

What is the difference between Devstral 2 2512 and GLM 4.7?

Devstral 2 2512 is developed by Mistral AI while GLM 4.7 is developed by OpenRouter. You can compare their actual outputs across 48 challenges on Rival to see how they differ in practice.

Which is better, Devstral 2 2512 or GLM 4.7?

It depends on your use case. Devstral 2 2512 and GLM 4.7 each have strengths in different areas. Rival lets you compare their real outputs side-by-side across 48 challenges so you can judge which fits your needs best.

How can I compare Devstral 2 2512 and GLM 4.7 on Rival?

This page shows a side-by-side comparison of Devstral 2 2512 and GLM 4.7 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

  • Devstral 2 2512 vs Gemini 4 ArgonLanded Oct 2026
  • GLM 4.7 vs Muse WonderLanded Oct 2026
  • Devstral 2 2512 vs GPT-4 TurboLanded Oct 2026
  • GLM 4.7 vs Step 5 PreviewLanded Oct 2026
  • Devstral 2 2512 vs Claude Haiku 5.5Landed Oct 2026
  • GLM 4.7 vs Ling 3.1 FlashLanded Oct 2026
  • Devstral 2 2512 vs Mistral Large 4Landed Oct 2026
  • GLM 4.7 vs GPT-6.1 SolLanded Sep 2026

Same lab, same size, long tail

  • Devstral 2 2512 vs Mistral Small 4Same lab
  • Devstral 2 2512 vs Step 5 PreviewSame size
  • GLM 4.7 vs Ox AlphaSame lab
  • GLM 4.7 vs OpenRouter Fusion · Budget (Jun 2026)Same lab
  • Devstral 2 2512 vs Qwen3.5 27BNew provider
  • Devstral 2 2512 vs Qwen3.5 35B A3BNew provider
  • Devstral 2 2512 vs Qwen3.5 397B A17BSame size
  • Devstral 2 2512 vs Qwen3.5 9BNew provider

Model pages

  • Devstral 2 251248 outputs, specs and price
  • GLM 4.771 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