Mistral Ministral 3 8B
Mistral's 2025 edge MoE: 8B params, 256K context, Apache 2.0, native tool calling.
About
Ministral 3 is Mistral AI's 2025 small-model line (3B, 8B, 14B), with the 8B as the balanced pick: 256K context, text and image inputs, native function calling and JSON mode, all under Apache 2.0. It targets on-device and edge deployment where a 70B+ model is impractical, with the vLLM parser integration making it drop-in for agent frameworks.
Limitations: small-model ceiling on complex reasoning — it won't match Mistral Large on hard tasks. Image support is narrower than dedicated vision models. The 14B is the stronger choice if VRAM allows.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Team-reported: strong efficiency on edge benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 44 (models); Trending score 74/100 from researcher assessment.