Dataconomy
  • News
    • Artificial Intelligence
    • Cybersecurity
    • DeFi & Blockchain
    • Finance
    • Gaming
    • Startups
    • Tech
  • Industry
  • Research
  • Resources
    • Articles
    • Guides
    • Case Studies
    • Whitepapers
    • AI Models Leaderboard
  • AI toolsNEW
  • Newsletter
  • + More
    • Glossary
    • Conversations
    • Events
    • About
      • Who we are
      • Contact
      • Imprint
      • Legal & Privacy
      • Partner With Us
Subscribe
No Result
View All Result
  • AI
  • Tech
  • Cybersecurity
  • Finance
  • DeFi & Blockchain
  • Startups
  • Gaming
Dataconomy
  • News
    • Artificial Intelligence
    • Cybersecurity
    • DeFi & Blockchain
    • Finance
    • Gaming
    • Startups
    • Tech
  • Industry
  • Research
  • Resources
    • Articles
    • Guides
    • Case Studies
    • Whitepapers
    • AI Models Leaderboard
  • AI toolsNEW
  • Newsletter
  • + More
    • Glossary
    • Conversations
    • Events
    • About
      • Who we are
      • Contact
      • Imprint
      • Legal & Privacy
      • Partner With Us
Subscribe
No Result
View All Result
Dataconomy
No Result
View All Result

Mistral Medium 3 destroys others in benchmarks

Mistral states its Medium 3 model performs at or above 90% of Anthropic's Claude Sonnet 3.7 and surpasses Meta's Llama 4 Maverick on benchmarks.

byEmre Çıtak
May 8, 2025
in Artificial Intelligence, News
Home News Artificial Intelligence
Share on FacebookShare on TwitterShare on LinkedInShare on WhatsAppShare on e-mail
Google Preferred Source

Mistral, a French AI startup, has released its latest AI model, Mistral Medium 3, which promises leading performance at a competitive price. The model is available through Mistral’s API, priced at $0.40 per million input tokens and $2 per million output tokens.

Mistral claims that Mistral Medium 3 performs “at or above” 90% of Anthropic’s more expensive Claude Sonnet 3.7 model across various benchmarks. It also outperforms recent open models, including Meta’s Llama 4 Maverick and Cohere’s Command A, on popular AI performance evaluations. To put this into perspective, a million tokens are equivalent to about 750,000 words, roughly 163,000 words longer than “War and Peace.”

Mistral Medium 3 is optimized for coding and STEM tasks and excels at multimodal understanding. The company states that clients in financial services, energy, and healthcare have been beta testing the model for use cases like customer service, workflow automation, and analyzing complex datasets. The model can be deployed on any cloud, including self-hosted environments with four GPUs and above.

Stay Ahead of the Curve!

Don't miss out on the latest insights, trends, and analysis in the world of data, technology, and startups. Subscribe to our newsletter and get exclusive content delivered straight to your inbox.

Mistral Medium 3 destroys others in benchmarks
Image: Mistral

In addition to Mistral’s API, where enterprise customers can work with Mistral to fine-tune the model, Mistral Medium 3 is available on Amazon’s Sagemaker platform as of Wednesday. It will soon be available on other hosts, including Microsoft’s Azure AI Foundry and Google’s Vertex AI platforms.

The launch of Mistral Medium 3 follows the release of Mistral Small 3.1 in March. Mistral has also teased the release of a much larger model in the coming weeks. On Wednesday, the company launched Le Chat Enterprise, a corporate-focused chatbot service that offers tools like an AI “agent” builder and integrates Mistral’s models with third-party services like Gmail, Google Drive, and SharePoint.

Le Chat Enterprise will soon support MCP, Anthropic’s standard for connecting AI assistants to the systems and software where data resides. Mistral has raised over €1.1 billion (roughly $1.24 billion) to date, with backing from VCs including General Catalyst. Its customers include BNP Paribas, AXA, and Mirakl.


Featured image credit

Tags: AIMistral

Related Posts

OpenAI limits ChatGPT 5.6 access to government-approved users first

OpenAI limits ChatGPT 5.6 access to government-approved users first

June 26, 2026
Apple to skip M6 Pro and Max chips and launch M7 in 2027

Apple to skip M6 Pro and Max chips and launch M7 in 2027

June 26, 2026
IBM unveils world’s first sub-1nm chip with new nanostack architecture

IBM unveils world’s first sub-1nm chip with new nanostack architecture

June 26, 2026
Apple raises prices across Macs, iPads and home devices

Apple raises prices across Macs, iPads and home devices

June 26, 2026
Nothing to launch entry-level Phone 4b on July 7

Nothing to launch entry-level Phone 4b on July 7

June 26, 2026
Xbox tests 15-character gamertags for Insider users

Xbox tests 15-character gamertags for Insider users

June 26, 2026

LATEST NEWS

OpenAI limits ChatGPT 5.6 access to government-approved users first

Apple to skip M6 Pro and Max chips and launch M7 in 2027

IBM unveils world’s first sub-1nm chip with new nanostack architecture

Apple raises prices across Macs, iPads and home devices

Nothing to launch entry-level Phone 4b on July 7

Xbox tests 15-character gamertags for Insider users

BEST AI MODELS LEADERBOARD

See the best AI models, ranked by intelligence, benchmark results, speed and token price. Find the most suitable LLMs, Text-to-Image, Image Editing, Text-to-Speech, Text-to-Video and Image-to-Video  artificial intelligence model for your tasks and business.

LATEST TOOLS

WatchMyCompetitor

TokkingHeads

Fellow.app

Octoparse

AnyToSpeech

Vrew

Fireflies

SpeedLegal

Teachable Machine

Unriddle

Dataconomy

COPYRIGHT © DATACONOMY MEDIA GMBH, ALL RIGHTS RESERVED.

  • About
  • Imprint
  • Contact
  • Legal & Privacy

Follow Us

  • News
    • Artificial Intelligence
    • Cybersecurity
    • DeFi & Blockchain
    • Finance
    • Gaming
    • Startups
    • Tech
  • Industry
  • Research
  • Resources
    • Articles
    • Guides
    • Case Studies
    • Whitepapers
    • AI Models Leaderboard
  • AI tools
  • Newsletter
  • + More
    • Glossary
    • Conversations
    • Events
    • About
      • Who we are
      • Contact
      • Imprint
      • Legal & Privacy
      • Partner With Us
No Result
View All Result
Subscribe

This website uses cookies to improve your experience. You can choose to accept or reject them. Visit our Privacy Policy.