Mistral Small 4
Mistral’s efficient hybrid model combining instruct, reasoning, and coding capabilities at very low API pricing.
Capability profile
- Input
- $0.15 / 1M tokens
- Output
- $0.60 / 1M tokens
- Cached input
- Unknown
- Context
- 256k tokens
- Free tier
- No / unknown
- API
- Available
- Subscription
- Mistral platform/API
- Efficient instruct
- reasoning
- and coding workflows where low price matters more than absolute flagship capability.
- Coding assistance
- agent backends
- structured outputs
- document workflows
- and low-cost reasoning tasks.
- Top-end frontier reasoning where larger premium models should still win.
- Workloads needing maximum quality regardless of cost or million-token context windows.
Small 4 is not the glamour entry, but its hybrid design and low price make it a practical model-tracker candidate for builders watching margins.
Mistral’s efficient hybrid model combining instruct, reasoning, and coding capabilities at very low API pricing.
Post Reboot take
Small 4 is not the glamour entry, but its hybrid design and low price make it a practical model-tracker candidate for builders watching margins.
Strengths
Best for: Efficient instruct, reasoning, and coding workflows where low price matters more than absolute flagship capability.
Weakest at: Top-end frontier reasoning where larger premium models should still win.
Best workflows: Coding assistance, agent backends, structured outputs, document workflows, and low-cost reasoning tasks.
Example use cases: Budget coding agents, workflow automation, document Q&A, extraction, and structured support tools.