Mistral Large 2
Available frontierMistral's flagship model for enterprise reasoning
Released July 24, 2024
- Context Window
- 128K
- TTFT
- N/A
- Speed
- N/A
- Max Output
- 8K
- Training Cutoff
- 2024-06
Last tested: 2026-01-15
About
Mistral Large 2 represents the pinnacle of European AI engineering. With 123 billion parameters and a 128K context window, it delivers frontier-level reasoning, multilingual excellence across 80+ languages, and native function calling. Built in Paris by the team that pioneered efficient AI architectures, it offers enterprise-grade performance at competitive pricing.
Capabilities
Pricing
- Input
- $2.00/M per 1M tokens
- Output
- $6.00/M per 1M tokens
Static pricing
Details
Mistral Large 2
Mistral Large 2 is the flagship model from Mistral AI, the Paris-based laboratory that has rapidly become Europe’s leading AI company. Released in July 2024, this 123-billion parameter model represents a significant leap in capability while maintaining the efficiency-first philosophy that defines Mistral’s approach.
Key Strengths
Multilingual Excellence: With native support for over 80 languages including French, German, Spanish, Italian, Portuguese, Arabic, Chinese, Japanese, Korean, and Hindi, Mistral Large 2 excels at cross-lingual tasks and serves global enterprises without the need for multiple specialized models.
Advanced Reasoning: The model demonstrates strong performance on complex reasoning benchmarks, competing directly with GPT-4 and Claude on mathematical, logical, and multi-step reasoning tasks.
Code Proficiency: Trained extensively on code, Mistral Large 2 handles software engineering tasks with precision, supporting over 80 programming languages with particular strength in Python, Java, JavaScript, TypeScript, and C++.
Function Calling: Native support for structured tool use enables reliable agent workflows and API integrations, with consistent JSON schema adherence.
European AI Leadership
Mistral AI was founded in 2023 by former researchers from Google DeepMind and Meta AI. In just over a year, the company has established itself as a credible alternative to American AI labs, demonstrating that world-class AI development can happen in Europe. Mistral Large 2 embodies this mission, offering frontier capabilities with transparent pricing and European data sovereignty options.
Best Use Cases
- Enterprise reasoning and analysis
- Multilingual content generation
- Complex code generation and review
- Agentic workflows requiring tool use
- Document understanding and summarization
Provider
Other Mistral AI Models
See all models- Mixtral 8x22B
- Sparse mixture-of-experts model with frontier performance
- Context
- 64K
- Speed
- N/A
- TTFT
- N/A
- Mistral Small
- Cost-efficient model for high-volume applications
- Context
- 128K
- Speed
- N/A
- TTFT
- N/A
- Codestral
- Purpose-built for code generation and software engineering
- Context
- 32K
- Speed
- N/A
- TTFT
- N/A
Get the signal, skip the noise.
Weekly digest of new models and provider updates across 40+ compute providers. Curated for AI builders who ship.