Devstral Small 1.1
Devstral Small 1.1 is an Apache 2.0 open-source agentic coding model with 46.8% on SWE-bench Verified, built jointly by Mistral AI and All Hands AI.
import { streamText } from 'ai'
const result = streamText({ model: 'mistral/devstral-small', prompt: 'Why is the sky blue?'})Playground
Try out Devstral Small 1.1 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.
Devstral Small 1.1
Ask Devstral Small 1.1 anything to try it out.
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.
| Provider |
|---|
Mistral AI | 128K | 0.4s | 33tps | $0.10/M | $0.30/M | — | — | 05/21/2025 |
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 | 0.8s | 26tps | $0.10/M | $0.30/M | — | — | | 12/09/2025 | |||||
| 256K | 0.5s | 57tps | $0.50/M | $1.50/M | — | — | | 12/02/2025 | |||||
| 256K | 0.3s | 85tps | $0.20/M | $0.20/M | — | — | +1 | | 12/02/2025 | ||||
| 128K | 0.3s | 140tps | $0.10/M | $0.10/M | — | — | | 10/16/2024 | |||||
| 32K | 0.4s | 215tps | $0.10/M | $0.30/M | — | — | | 09/17/2024 | |||||
| $0.10/M | — | — | | 12/11/2023 |
About Devstral Small 1.1
Released May 21, 2025, Devstral Small 1.1 was built as a collaboration between Mistral AI and All Hands AI, the team behind the OpenHands agent scaffold. That partnership shaped Devstral Small 1.1's design for resolving real GitHub issues, giving it practical competency in the messy, context-dependent bugs that appear in production repositories.
At 46.8% on SWE-bench Verified, Devstral Small 1.1 outperformed previous open-source models by more than six percentage points at release. Under the OpenHands scaffold, it exceeded larger models like Deepseek-V3 (671B) and Qwen3 (232B). These results reflect efficient use of its parameter budget rather than scale.
Devstral Small 1.1 integrates with OpenHands, SWE-Agent, Ollama, LM Studio, and other tooling ecosystems.
What To Consider When Choosing a Provider
- Configuration: Devstral Small 1.1's Apache 2.0 license makes it among the more accessible open-source coding agents available through AI Gateway.
- 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 Devstral Small 1.1
Best For
- OpenHands or SWE-agent backend: Teams using these scaffolds need a compact backend model
- GitHub issue resolution prototypes: Early-stage exploration of automated issue-resolution pipelines
- Apache 2.0 licensing requirement: Projects that mandate an Apache 2.0 open-weight license
Consider Alternatives When
- Higher SWE-bench accuracy: You need a stronger score (consider Devstral 2)
- Multimodal input capabilities: Your agent requires image or other non-text input
- Reasoning-first approach: Your task needs reasoning rather than agentic code execution
Conclusion
Devstral Small 1.1 brought agentic software engineering to a compact open-weight model. Devstral Small 1.1's GitHub-issue training, Apache 2.0 license, and broad deployment ecosystem make it a practical choice for teams exploring AI-assisted code maintenance.