Mistral Small 3.1 24B Instruct

Building upon Mistral Small 3 (2501), Mistral Small 3.1 (2503) adds state-of-the-art vision understanding and enhances long context capabilities up to 128k tokens without compromising text performance. With 24 billion parameters, this model achieves top-tier capabilities in both text and vision tasks.

Benchmark results

Benchmark Score Tags Source
GPQA 46.0% self-reported llm-stats link →
HumanEval 88.4% self-reported llm-stats link →
MATH 69.3% self-reported llm-stats link →
MBPP 74.7% self-reported llm-stats link →
MMLU 80.6% self-reported llm-stats link →
MMLU-Pro 66.8% self-reported llm-stats link →
MMMU 59.3% self-reported llm-stats link →
SimpleQA 10.4% self-reported llm-stats link →
TriviaQA 80.5% self-reported llm-stats link →