Skip to content

Devstral 2 2512

Devstral 2 is a state-of-the-art open-source model by Mistral AI specializing in agentic coding.

Pricing

Input
$0.42per 1M tokens
Output
$2.10per 1M tokens
Cache read
$0.042per 1M tokens
Max output
209,715tokens

OpenRouter's price plus OpenRouter's 5% platform fee. Deference adds no markup.

opencode.json

{
  "provider": {
    "deference": {
      "npm": "@ai-sdk/openai-compatible",
      "options": { "baseURL": "https://deference.si/v1", "apiKey": "{env:DEFERENCE_API_KEY}" },
      "models": { "mistralai/devstral-2512": { "limit": { "context": 262144, "output": 209715 } } }
    }
  }
}
OpenCode, set up for Devstral 2 2512

Overview

It is a 123B-parameter dense transformer model supporting a 256K context window.

Find us on