LangMart: Mistral: Devstral Small 2505
Model Overview
| Property | Value |
|---|---|
| Model ID | openrouter/mistralai/devstral-small-2505 |
| Name | Mistral: Devstral Small 2505 |
| Provider | mistralai |
| Released | 2025-05-21 |
Description
Devstral-Small-2505 is a 24B parameter agentic LLM fine-tuned from Mistral-Small-3.1, jointly developed by Mistral AI and All Hands AI for advanced software engineering tasks. It is optimized for codebase exploration, multi-file editing, and integration into coding agents, achieving state-of-the-art results on SWE-Bench Verified (46.8%).
Devstral supports a 128k context window and uses a custom Tekken tokenizer. It is text-only, with the vision encoder removed, and is suitable for local deployment on high-end consumer hardware (e.g., RTX 4090, 32GB RAM Macs). Devstral is best used in agentic workflows via the OpenHands scaffold and is compatible with inference frameworks like vLLM, Transformers, and Ollama. It is released under the Apache 2.0 license.
Description
LangMart: Mistral: Devstral Small 2505 is a language model provided by mistralai. This model offers advanced capabilities for natural language processing tasks.
Provider
mistralai
Specifications
| Spec | Value |
|---|---|
| Context Window | 128,000 tokens |
| Modalities | text->text |
| Input Modalities | text |
| Output Modalities | text |
Pricing
| Type | Price |
|---|---|
| Input | $0.06 per 1M tokens |
| Output | $0.12 per 1M tokens |
Capabilities
- Frequency penalty
- Max tokens
- Min p
- Presence penalty
- Repetition penalty
- Response format
- Seed
- Stop
- Temperature
- Top k
- Top p
Detailed Analysis
Devstral Small 2505 (May 2025 release) represents an updated version of Devstral Small with improvements in code understanding and tool usage efficiency. This release incorporates learnings from production usage patterns and includes enhanced training on 2025 framework updates (React 19, Next.js 15, Python 3.13, TypeScript 5.5+). The model features improved context retention across multi-turn development sessions, better understanding of modern DevOps practices, and refined tool calling sequences for common workflows. Devstral Small 2505 maintains the fast inference characteristics of the original Small variant while addressing edge cases in dependency resolution, type inference, and error handling. The May 2025 update also includes better support for emerging development tools, AI-native development patterns, and enhanced ability to work with monorepo structures and microservices architectures.