Mistral Large 3 675B is Mistral AI's flagship language model featuring advanced rope scaling and Eagle speculative decoding. Delivers exceptional performance across reasoning, coding, and multilingual tasks.
Added Dec 25, 2025
Context Window
262.1K
Max Output
256.0K
Avg output tokens (7d)
552 tokens
Input Price (Auto)
$1.00/1M
Output Price (Auto)
$3.00/1M
Cache Read (Auto)
$0.50/1M
Capabilities
Benchmarks
Benchmarks
Performance metrics and benchmarks
Sourced from Artificial Analysis.
Intelligence Index
9.7
Coding Index
20.1
Agentic Index
2.4
Agentic work
AutomationBench-AA
Workflow automation with guardrail penalties
1.5%
Better than 21% of models compared
AA-Briefcase
Agentic knowledge work (Elo)
246 Elo
Better than 15% of models compared
GDPval-AA v2
Economically valuable tasks (Elo)
588 Elo
Better than 25% of models compared
Document reasoning
GDP.pdf
Professional PDF reasoning: all-pass rate
1.2%
Better than 13% of models compared
AA-LCR v1.1
Long context reasoning with updated grading
36.0%
Better than 36% of models compared
Reasoning
HLE
Humanity's Last Exam
4.2%
Better than 17% of models compared
IFBench
Instruction-following benchmark
36.2%
Better than 28% of models compared
CritPt
Research-level physics reasoning
0.0%
Coding
Terminal-Bench v4.0
Practical coding and terminal tasks
0.0%
Better than 17% of models compared
SciCode
Python programming for scientific computing
36.6%
Better than 19% of models compared
LiveCodeBench
Contamination-free coding benchmark
46.5%
Better than 53% of models compared
Math
AIME 2025
American Invitational Mathematics Examination 2025
38.0%
Better than 39% of models compared
Knowledge
MMLU-Pro
Professional and academic subject knowledge
80.7%
Better than 70% of models compared
AA-Omniscience Accuracy
Proportion of correctly answered questions
24.9%
AA-Omniscience Hallucination Rate
Rate of incorrect answers among non-correct responses
86.0%
Legacy benchmarks
GPQA Diamond (legacy)
Graduate-level scientific reasoning
68.0%
Better than 45% of models compared
Terminal-Bench Hard (legacy)
Agentic coding and terminal use
15.9%
Better than 53% of models compared
T²-Bench Telecom (legacy)
Conversational AI agents in dual-control scenarios
24.6%
Better than 27% of models compared
AA-LCR (unversioned / legacy)
Long context reasoning evaluation
36.0%
Better than 36% of models compared
GDPval-AA (unversioned / legacy)
Economically valuable tasks
4.4%
Last updated Sep 13, 2026
Artificial AnalysisProviders
Auto routing is available for this model. Explicit provider selection is not available.
Loading provider options…
Related text models
Compare Mistral Large 3 675B with similar models from the same provider or model family.
Mistral Small 3 24B (2501)
mistralai/mistral-small-24b-instruct-2501Mistral Small 3 24B (2501) hosted by IONOS in Berlin, Germany. Zero data retention.
Devstral 2 123B
mistralai/devstral-2-123b-instruct-2512Devstral 2 123B is a 123 billion parameter model from Mistral AI optimized for coding and development tasks. Features advanced reasoning capabilities for software engineering workflows.
Ministral 3 14B
mistralai/ministral-14b-instruct-2512Ministral 3 14B is a balanced model in the Ministral 3 family, designed for edge deployment. A powerful, efficient language model with vision capabilities, fine-tuned for instruction tasks. Features multilingual support, strong system prompt adherence, and native function calling. Apache 2.0 licensed.
Mistral Large 2411
mistralai/mistral-largeUpgrade to Mistral's flagship model. It is fluent in English, French, Spanish, German, and Italian, with high grammatical accuracy, with a long context window.
Mistral Nemo
mistralai/mistral-nemo-instruct-240712B parameter model with multilingual support.
Mistral Small 4 119B Thinking
mistralai/mistral-small-4-119b-2603:thinkingMistral Small 4 with reasoning enabled (reasoning_effort=high). A hybrid MoE model with deep step-by-step reasoning for complex prompts, coding, and multi-step problem solving.