Mistral Small
A compact, efficient open-weights model released under Apache-2.0, ideal for self-hosting and edge deployment. Delivers solid general capability at very low cost.
At a glance
Released
Jul 22, 2026
Provider
Mistral AI
Context window
128,000 tokens
Max output
32,000 tokens
Modalities
Text + image → text
Knowledge cutoff
April 2026
API model ID
mistral-small-latestLicense
Apache-2.0
Pricing
Standard
Input$0.10 / 1M
Output$0.30 / 1M
* Rates are from a third-party aggregator — verify before relying on them. A typical 1,500-token prompt with a 650-token answer costs ≈ $0.0003 on the standard tier. Run your own prompt through the calculator →
What's new vs the previous model
Compared withMistral Small 3 (October 2025)
- Better instruction following and efficiency than Small 3
- Shorter context and lower ceiling than Mistral Large
Best use cases
Self-hosting
Runs on modest hardware under a permissive Apache-2.0 license.
Edge and on-prem
Powers local assistants where data cannot leave the premises.
High-volume tasks
Handles classification and drafting cheaply at scale.
Strengths & limitations
Strengths
- Fully open weights
- Low cost and resource needs
- Permissive license
Limitations
- Lower capability ceiling
- Smaller context window
How to test it
- No-code: open chat.mistral.ai and pick Mistral Small from the model switcher — no API key needed.
- Playground: sign in to La Plateforme (the Mistral console) and run prompts in the workbench, then copy the generated code.
- API: create a key in La Plateforme (the Mistral console) and call the API (example below).
curl https://api.mistral.ai/v1/chat/completions \
# OpenAI-compatible endpoint
-H "Authorization: Bearer $API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "mistral-small-latest",
"messages": [{"role": "user", "content": "Explain tokenization in one paragraph."}]
}'Where to get access
Direct API
Pay-as-you-go access via La Plateforme (the Mistral console).
La Plateforme (the Mistral console) ↗