Home/Models/Mistral Medium 3
Mistral AI logo

Mistral Medium 3

by Mistral AI · Released May 2025

Open SourceMultimodal
41.1
avg score
Rank #141
Compare
Better than 39% of all models
Context
131K tokens (~66 books)
Input $/1M
$0.40
Output $/1M
$2.00
Type
multimodal
License
Open Source
Benchmarks
4 tested
Data updated today
About

Mistral Medium 3 is a high-performance enterprise-grade language model designed to deliver frontier-level capabilities at significantly reduced operational cost. It balances state-of-the-art reasoning and multimodal performance with 8× lower cost...

Tested on 4 benchmarks with 40.0% average. Top scores: MATH level 5 (81.6%), GPQA diamond (46.0%), OTIS Mock AIME 2024-2025 (32.1%).

Looking for similar performance at lower cost?
Llama 3 8B Instruct scores 41.7 (101% as good) at $0.03/1M input · 93% cheaper
Capabilities
math
38.0
#108 globally
knowledge
46.0
#121 globally
Benchmark Scores
Compare All
Tested on 4 benchmarks · Ranked across 2 categories
Score Distribution (all 233 models)
0255075100
▲ You are here
MATH level 5

Competition-level math from AMC, AIME, and olympiad problems. Level 5 is the hardest tier, requiring creative problem-solving.

81.6
OTIS Mock AIME 2024-2025

Mock AIME (American Invitational Mathematics Exam) problems from OTIS. Tests mathematical competition performance.

32.1
FrontierMath-2025-02-28-Private

Original research-level math problems created by professional mathematicians. Problems are unpublished and cannot be memorized.

0.3
GPQA diamond

Graduate-level science questions written by PhD experts. Diamond subset contains questions where experts disagree, testing deep understanding.

46.0
Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Links
Documentation
Community
BenchGecko API
mistral-medium-3
Specifications
  • Typemultimodal
  • Context131K tokens (~66 books)
  • ReleasedMay 2025
  • LicenseOpen Source
  • StatusActive
  • Cost / Message~$0.003
Available On
Mistral AI logoMistral AI$0.40
Categories
Share & Export
Tweet
Mistral Medium 3 is an open-source multimodal AI model by Mistral AI, released in May 2025. It has an average benchmark score of 41.1. Context window: 131K tokens.