MistralHosted on Luminet

Devstral 2

mistral/devstral-2

Try in Playground

Coding-agent fine-tune. Apache 2.0, 24B dense, deploys on a single H100. Tops Aider Polyglot among <30B models.

Context window

128Ktokens

Input price

$0.10/ 1M tokens

Output price

$0.30/ 1M tokens

P50 latency

92ms

Quick start

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.luminet.ai/v1",
  apiKey: process.env.LUMINET_API_KEY,
});

const resp = await client.chat.completions.create({
  model: "mistral/devstral-2",
  messages: [{ role: "user", content: "Hello!" }],
});

Capabilities

textcode

Throughput: 580 tok/s · Released 2026-03