Mistral Small 3
Provider: Mistral AI EU
Overview
Mistral Small 3 is a 24B parameter budget model from Mistral AI, built for low-latency common AI workloads under an Apache 2.0 license.
Strengths
- 24B parameter size
- Apache 2.0 license
- Built-in JSON mode support
- 32768 token context window
Best use cases
- Low-latency text processing tasks
- Structured JSON generation workflows
- Cost-sensitive internal tooling
Pricing position
Input costs $0.04/1M and output costs $0.064/1M, pricing it slightly below Gemma 3 4B and cheaper on output than Llama 3.1 8B.
Caveats
Context window is limited to 32768 tokens, making Mistral Small 4 or Ministral 3 8B better picks for long documents.
Notes drafted by an LLM from the figures on this page (2026-11-15) and rewritten when the price changes.
Key facts
- Price basis
- Median third-party host price (no first-party API)
- Max output
- 16K tokens
- Knowledge cutoff
- 2023-10-31
- Listed on OpenRouter
- 2025-01-30
- Tracked since
- 2026-05-08
- Last updated
- 2026-11-15
Features
Price history
- 2026-11-15 Price change $0.05 / $0.08 → $0.04 / $0.064
Monthly cost estimate
| Scale | Input (M) | Output (M) | Monthly |
|---|---|---|---|
| 0.1M (hobby) | 0.1 | 0.05 | $0.01 |
| 1M (small app) | 1 | 0.5 | $0.07 |
| 10M (mid scale) | 10 | 5 | $0.72 |
| 100M (large) | 100 | 50 | $7.20 |
※ Cache and batch discounts not factored in. Detailed calculator
Official resources
Other models from Mistral AI
Similar-priced models
Source: OpenRouter API (mistralai/mistral-small-24b-instruct-2501)