Devstral Small 2505
Mistral's Devstral Small 2505 is a chat-based, agentic large language model designed for software engineering tasks, exceling at using tools to explore codebases and editing multiple files. It achieves a notable score of 46.8% on the SWE-Bench Verified benchmark.
Input
Output
Context
131K
Max Output
128K
Parameters
23.6B
Technical Specifications
Model TypeCode
Context Window131,072 tokens
Max Output Tokens128,000 tokens
Parameters23.6B
Release DateMay 12, 2025
Training CutoffMar 1, 2025
Licenseapache-2.0
Open SourceYes
Input Modalities
Text
Output Modalities
Text
Capabilities
Estimated GPU Requirements for Devstral Small 2505
Estimates based on INT8 quantization. Actual requirements vary by framework and configuration.
More Models by Mistral
Voxtral Mini Realtime Charente 2606 Solutions33K context · Open weightsLabs Leanstral 1.5.1262K context · Open weightsVoxtral Mini 4B Realtime 2602131K context · 4.4B · Open weightsMistral Small 2506131K context · Open weightsMagistral Small 2509131K context · 24B · Open weightsMinistral 3 8B Instruct 2512 BF16262K context · 8.9B · Open weights
Related Tools
Data sourced from official provider APIs and documentation
Last updated: Aug 8, 2026
Your AI stack shouldn't stand still.
Every month new models become cheaper, faster, and more capable. Inferbase ensures your application automatically benefits without changing a single API call.