Skip to main content
Mistral AI builds models tuned for efficiency: frontier-adjacent capability on a throughput-oriented tier. Mistral Medium 3.1 delivers translation and content generation across 50+ languages, with a 130K-token context window and multimodal text-plus-image input.
Available in Ollang Workflows as mistralai/mistral-medium-3.1 — Document Translation, Subtitle Translation, and AI Dubbing. See the Translation catalog.
Migration notice. Mistral has deprecated Medium 3.1 and Medium 3.2 in favor of Mistral Medium 3.5 (mistral-medium-3505), released April 2026 and described as a frontier-class multimodal model optimized for agentic and coding use cases. Retirement dates for the 3.x series run through mid-2026. Plan a provider change for Workflows that depend on this tier.

​
Model and coverage

​
Key capabilities

  • Contextual and idiomatic fidelity across dozens of languages, with particular strength on European pairs.
  • Efficiency — engineered to deliver frontier-adjacent results on a model tier tuned for throughput rather than maximum capability.
  • Multimodal input for documents that mix text and images.
  • 130K context — enough for most single documents and full subtitle files in one pass.

​
Where it fits in a workflow

Mistral Medium 3.1 is a strong subtitle translation workhorse for European language pairs. Its advantage over the pure NMT engines is that it is an instruction-following LLM: it will apply style guidance and reason about ambiguous source text in a way DeepL and Amazon will not.

​
Trade-offs

  • Deprecated upstream. See the migration notice above.
  • 50+ languages is solid but well short of the 100+ providers; verify coverage for non-European pairs.
  • Not a frontier reasoning tier. For quality-critical, nuance-heavy content the Claude 5, GPT-5.6, and Gemini Pro tiers pull clearly ahead.

​
Reference