Mistral Small
Mistral Small (September 1, 2024) is a 22B mid-tier model with improved reasoning, alignment, and code at about 80% lower list pricing than its predecessor in Mistral AI's September 2024 table.
import { streamText } from 'ai'
const result = streamText({ model: 'mistral/mistral-small', prompt: 'Why is the sky blue?'})Playground
Try out Mistral Small by Mistral AI. Usage is billed to your team at API rates. Free users (those who haven't made a payment) get $5 of credits every 30 days.
About Mistral Small
Released September 1, 2024, the Mistral Small update was primarily a pricing event, with an approximately 80% reduction in input and output token pricing compared to the prior release. Simultaneous capability improvements in human alignment, reasoning, and code generation made this a significant upgrade rather than just a price cut.
At 22B parameters, Mistral Small sits between Mistral AI NeMo (12B) and Mistral AI Large 2 in the lineup Mistral AI described in September 2024. Mistral Small targets translation, summarization, and sentiment analysis.
The pricing cut plus capability improvements made Mistral Small a common pick for production workloads that previously ran larger, more expensive Mistral AI endpoints.
Providers
Route requests across multiple providers. Copy a provider slug to set your preference. Visit the docs for more info. Using a provider means you agree to their terms, listed under Legal.
P50 throughput on live AI Gateway traffic, in tokens per second (TPS). Visit the docs for more info.
P50 time to first token (TTFT) on live AI Gateway traffic, in milliseconds. View the docs for more info.
Direct request success rate on AI Gateway and per-provider. Visit the docs for more info.
More models by Mistral AI
| Model |
|---|
| 256K | 0.4s | 12tps | $0.40/M | $2.00/M | — | — |  | 12/09/2025 | |||||
| 256K | 0.4s | 47tps | $0.50/M | $1.50/M | — | — | | 12/02/2025 | |||||
| 256K | 0.2s | 126tps | $0.20/M | $0.20/M | — | — | | 12/01/2025 | |||||
| 256K | 0.4s | 29tps | $0.10/M | $0.30/M | — | — | | 05/07/2025 | |||||
| 128K | 0.2s | 257tps | $0.10/M | $0.10/M | — | — | | 10/01/2024 | |||||
| $0.10/M | — | — | | 12/11/2023 |
What To Consider When Choosing a Provider
- Configuration: At 22B parameters, Mistral Small sits between Mistral AI NeMo 12B and Mistral AI Large 2 as a mid-point in the lineup, per Mistral AI's September 2024 post.
- Zero Data Retention: AI Gateway supports Zero Data Retention for this model via direct gateway requests (BYOK is not included). To configure this, check the documentation.
- Authentication: AI Gateway authenticates requests using an API key or OIDC token. You do not need to manage provider credentials directly.
When to Use Mistral Small
Best For
- Translation and summarization at scale: High-volume translation, summarization, and sentiment analysis workloads
- 22B mid-tier production: Workloads that need competitive per-token pricing
- Mid-point between NeMo and Large: Applications where this capability fit is right
- Cost optimization: For workloads previously using larger Mistral AI models
Consider Alternatives When
- Higher reasoning capability: For workloads that need more depth (consider Mistral AI Large 3 or Magistral)
- Tekken tokenizer compression: You need multilingual compression advantages (consider Mistral AI NeMo)
- Vision input: Your workload needs image input (consider Pixtral models)
Conclusion
Mistral Small's September 1, 2024 update cut list pricing and improved alignment, reasoning, and code quality in Mistral AI's post. An 80% cost reduction shipped alongside those capability changes. Mistral Small remains a fit for teams that want a 22B mid-tier model through AI Gateway.
Frequently Asked Questions
What are the input and output prices for Mistral Small?
This page lists the current rates. Multiple providers can serve Mistral Small, so AI Gateway surfaces live pricing rather than a single fixed figure.
How many parameters does Mistral Small have?
22 billion parameters.
Where does Mistral Small sit in the model lineup?
Between Mistral AI NeMo (12B) and Mistral AI Large 2, described as a convenient mid-point for enterprise use cases.
What capability improvements came with the September 1, 2024 release?
Improved human alignment, stronger reasoning, and better code generation compared to the prior Mistral Small version.
What tasks is Mistral Small well-suited for?
Translation, summarization, and sentiment analysis. Mistral AI positions Mistral Small for tasks that don't need Mistral AI Large-scale breadth.
How does Mistral Small compare to Mistral AI NeMo?
Mistral Small is larger (22B vs 12B), more capable, and more expensive. Mistral AI NeMo has the Tekken tokenizer's compression advantages for code and non-Latin scripts. The choice depends on capability requirements and token efficiency needs.