Mistral AI
Mistral Large 2512
The European flagship — open-weights, EU-hosted options.
Open weightstextimagefileREL 2025-12-01
Compare this model →Verified 2026-06-12
Overview
Mistral Large 2512 is Mistral's flagship with open weights and strong function calling, at $0.50/$1.50 per 1M tokens with a 262K context. The go-to for EU data-residency requirements and self-hosted European deployments.
Benchmark scores
Methodology →Intelligence23
Codingn/a
Reasoningn/a
Mathn/a
Sources and verification
- Intelligence: Artificial Analysis Intelligence Index · checked 2026-06-12
- Catalogue and pricing: OpenRouter model record · checked 2026-06-12
Strengths and weaknesses
Strengths
- Open weights + EU hosting
- Good function calling
- Cheap
Trade-offs
- Behind US/Chinese frontier on benchmarks
- 262K context
Reach for this model when
EU-compliance deployments and self-hosted European stacks.
Run something like it locally
Designed for self-hosting — vLLM on multi-GPU, or quantised via llama.cpp.
