O

LangMart: Mistral: Devstral Medium

Openrouter
131K
Context
$0.4000
Input /1M
$2.00
Output /1M
N/A
Max Output

LangMart: Mistral: Devstral Medium

Model Overview

Property Value
Model ID openrouter/mistralai/devstral-medium
Name Mistral: Devstral Medium
Provider mistralai
Released 2025-07-10

Description

Devstral Medium is a high-performance code generation and agentic reasoning model developed jointly by Mistral AI and All Hands AI. Positioned as a step up from Devstral Small, it achieves 61.6% on SWE-Bench Verified, placing it ahead of Gemini 2.5 Pro and GPT-4.1 in code-related tasks, at a fraction of the cost. It is designed for generalization across prompt styles and tool use in code agents and frameworks.

Devstral Medium is available via API only (not open-weight), and supports enterprise deployment on private infrastructure, with optional fine-tuning capabilities.

Description

LangMart: Mistral: Devstral Medium is a language model provided by mistralai. This model offers advanced capabilities for natural language processing tasks.

Provider

mistralai

Specifications

Spec Value
Context Window 131,072 tokens
Modalities text->text
Input Modalities text
Output Modalities text

Pricing

Type Price
Input $0.40 per 1M tokens
Output $2.00 per 1M tokens

Capabilities

  • Frequency penalty
  • Max tokens
  • Presence penalty
  • Response format
  • Seed
  • Stop
  • Structured outputs
  • Temperature
  • Tool choice
  • Tools
  • Top p

Detailed Analysis

Devstral Medium is a 24B parameter model designed for agentic coding workflows that span entire codebases. Unlike Codestral (single-file operations), Devstral excels at complex software development tasks requiring navigation across multiple files, understanding project structure, and coordinating tool usage. It leverages advanced function calling to interact with development tools, version control systems, linters, test runners, and debugging utilities. Devstral Medium achieves 61.6% on SWE-Bench Verified, demonstrating state-of-the-art performance on real-world software engineering challenges. The model understands architectural patterns, dependency relationships, and can perform multi-step refactoring across an entire project. It's optimized for semi-automated software development where AI assists with complex tasks like feature implementation, bug investigation spanning multiple modules, code migration, and architectural refactoring. The 'medium' designation indicates balanced performance between speed and capability for production development workflows.