Devstral Medium

Provider: Mistral AI EU

No longer listed on OpenRouter since 2026-05-31. The figures below are the last recorded values.
Input price
$0.40
/ 1M tokens
Output price
$2.00
/ 1M tokens
Cache price
$0.04
/ 1M tokens (read)
Context
131K tokens
Max output Unknown

Overview

Mistral AI's Devstral Medium is a capable mid-tier model targeting developer-centric workloads that require sophisticated structured interactions. It offers a very large context window, well-suited for processing extensive information and complex instructions.

Strengths

  • Supports extensive inputs and long conversations with its 131,072 token context window.
  • Enables reliable automation and tool orchestration through native function-calling capabilities.
  • Facilitates precise data extraction and structured output generation via dedicated JSON-mode.
  • Features a unique $0.04/M token cache read price, potentially reducing costs for repeated data access.

Best use cases

  • Automating complex multi-step workflows requiring interactions with external APIs or tools.
  • Summarizing or analyzing extensive documents, reports, or large codebases effectively.
  • Generating highly structured data outputs for seamless integration into databases or applications.
  • Developing intelligent agents that can reliably parse and respond with structured information.

Pricing position

Devstral Medium is positioned as a higher-tier offering from Mistral AI, priced at $0.40/M for input and $2.00/M for output tokens, making it significantly more expensive than models like Mistral Small 3 ($0.05/$0.08). While its input cost matches peers like GPT-4.1 Mini and Llama 3.1 70B, its output price of $2.00/M is notably higher than GPT-4.1 Mini's $1.60/M and Llama 3.1 70B's $0.40/M.

Caveats

The maximum output token limit for Devstral Medium is currently unspecified, which may be a consideration for tasks requiring very long generations. There is no free tier available, meaning all usage will incur costs. For simpler or more cost-sensitive tasks not requiring extensive context or its specific capabilities, other Mistral models like Mistral Small 3 might offer a more economical solution.

Notes drafted by an LLM from the figures on this page (2026-05-21) and rewritten when the price changes.

Key facts

Max output
Unknown
Tracked since
2026-05-08
Last updated
2026-05-31

Features

Function callingJSON output

Price history

  • 2026-05-31 No longer listed

Monthly cost estimate

Scale Input (M) Output (M) Monthly
0.1M (hobby) 0.1 0.05 $0.14
1M (small app) 1 0.5 $1.40
10M (mid scale) 10 5 $14.00
100M (large) 100 50 $140.00

※ Cache and batch discounts not factored in. Detailed calculator

Official resources

Other models from Mistral AI

Similar-priced models

Source: OpenRouter API