Mistral Small 4
Apache-licensed open model unifying reasoning, vision and coding.
Mistral Small 4 unifies the capabilities of several Mistral models into a single 119B system, combining reasoning, vision and coding under an Apache 2.0 license.
- Provider
- Mistral AI
- Type
- Language model
- Released
- Mar 16, 2026
- Context window
- 262K tokens
- Max output
- 210K tokens
- Price
- $0.15 input / $0.60 output per 1M tokens
- Input
- text, image
- Output
- text
- Open weights
- Yes
- License
- Apache 2.0
Best for
- Local & on-device
- High volume / low cost
- Vision
- Extraction & classification
- Real-time / low latency
Strengths
- Apache 2.0
- $0.02 cost per task
- Fast
More from Mistral AI
- Mistral Large 4 — Mistral AI. Mistral’s 1T-parameter multimodal MoE flagship (preview; weights promised).
- Mistral Medium 3.5 — Mistral AI. Fast enterprise-grade multimodal model.
- Voxtral Small — Mistral AI. Best open-weight transcription model — also understands audio.