Skip to content

Magistral Medium 1.2

Released 2025-09 · reasoning · 128K tokens · 4 benchmarks · Open weight

Editorial notes

Primer modelo de razonamiento de Mistral. Entrenamiento con RL puro (sin distillation). Multimodal en v1.2.

Spec sheet

Company
Mistral AI
Country
FR
Type
reasoning
Release
2025-09
Context
128K tokens
License
apache-2.0
Quants
BF16, Q8_0, Q5_K_M, Q4_K_M
Pricing (mistral)
$0.4/$1.2/M
Slug
magistral-medium-1-2

Benchmarks (4)

Cite this model
BibTeX · APA

BibTeX

@misc{frontier-magistral-medium-1-2,
  title  = {Magistral Medium 1.2},
  author = {{Mistral AI}},
  year   = {2025},
  note   = {Frontier Benchmarks AI atlas. Accessed 2026-05-08},
  url    = {https://frontierbenchmarks.com/models/magistral-medium-1-2}
}

APA

Mistral AI (2025). Magistral Medium 1.2 [Large language model]. Frontier Benchmarks AI. Retrieved 2026-05-08, from https://frontierbenchmarks.com/models/magistral-medium-1-2

Citation reflects the atlas page, not the original model paper. For the paper, see the "Resources" section above.