Devstral Small 1.1Open Weight

Released

Devstral Small 1.1 is an AI model released by Mistral on Thursday, Jul 10 2025, 20 days after Mistral Small 3.2. It is an open-weight model — the trained weights are available to download and run. It is a 24B parameterThe learned values that shape how a model responds. B means billion and T means trillion. For mixture-of-experts models, the total count is larger than the number active for each token. Parameter count alone does not measure quality. model with a 128k token context windowHow much information the model can work with in one request, measured in tokens — pieces of words or other data. This includes your question, attachments, conversation and reply. Limits can vary by model variant and provider.. Benchmark results (shown below) cover SWE-Bench Verified.

Benchmarks

SWE-Bench VerifiedCoding — Real coding tasks pulled from open-source projects — the AI has to find and fix actual bugs. A human-checked version of the original SWE-Bench. Higher is better.
53.6%
#67 of 87

Compare Devstral Small 1.1 with

Devstral Small 1.1

About

Devstral Small 1.1 was released on July 10, 2025 as an update to Mistral's 24B software-engineering model developed with All Hands AI. It followed the May Devstral Small release and targeted agents that investigate repository issues, edit files and use tools to check their work. Mistral reported 53.6% on SWE-bench Verified using the OpenHands scaffold.

The model retained a 128K-token context and downloadable Apache 2.0 weights. Its evaluation measured the model operating inside a specific agent environment, rather than standalone code completion. That distinction helps explain its place beside Codestral: Devstral was trained for sustained repository work, while Codestral also served latency-sensitive editor completions.

Frequently asked questions

Devstral Small 1.1 was released by Mistral on Thursday, Jul 10 2025.

All Mistral releases

45 tracked