Benchmark Scores
82/100MMLU84
HumanEval84.5
Math74.5
Reasoning82.8
Coding84
About Mistral Large 2
Mistral Large 2 is Mistral AI's flagship model with 123B parameters. It offers strong multilingual capabilities, function calling, and reasoning with a 128K context window, positioning it as a competitive European alternative.
Strengths
- +Strong multilingual (12+ languages)
- +Good function calling
- +Large context window
- +Competitive European alternative
Weaknesses
- -Below frontier on hardest benchmarks
- -Less ecosystem than OpenAI/Anthropic
- -Code generation not top-tier
Pricing
Per 1M tokens
Input$2.00
Output$6.00
Quick Facts
- Context Window
- 128K
- Parameters
- 123B
- Release Date
- 2024-07-24
- Category
- Mid-Tier
Compare Models
See how Mistral Large 2 stacks up against other models