Mistral Large 3
MistralFlagship
Tool UseVisionStructured Output
About this model
Mistral Large 3 2512 is Mistral’s most capable model to date, featuring a sparse mixture-of-experts architecture with 41B active parameters (675B total), and released under the Apache 2.0 license.
Performance Tier
Flagship
Mistral Large 3 is a flagship model from Mistral : the most capable in their lineup.
Best-in-class model from this provider. Highest performance across benchmarks, ideal for demanding tasks.
Pricing
This model is included in Elosia plans| Type | per 1M tokens |
|---|---|
| Input (prompt) | $0.500 |
| Output (completion) | $1.50 |
| Cache read | $0.050 |
Capabilities
Context Length262K
Max Output Tokens—
TokenizerMistral
Inputtext, image
Outputtext
Release DateDecember 1, 2025
Benchmarks
General Intelligence
MMLU
85.5%
GPQA Diamond
43.9%
Mathematics
MATH-500
93.6%
Programming
HumanEval
92%
SWE-bench Verified
52.8%
Reasoning
IFEval
85%
Recommended Use Cases
CodingTranslationMathematicsGeneral Chat
Strengths
- Excellent math reasoning (MATH-500 93.6%)
- Strong code generation (HumanEval 92%)
- Fast inference with MoE architecture (24B active / 387B total)
- EU-based company (data sovereignty)
Limitations
- Weaker on graduate-level science reasoning (GPQA 43.9%)
- Less capable than top-tier models on complex agentic tasks
Resources
This model may use your data for training
