✍️Writing & Content58🎨Image Generation71🎬Video & Animation120🎵Audio & Music100💬Chatbots & Assistants109💻Coding & Development444📈Marketing & SEO197Productivity401🎯Design & UI/UX120📊Data & Analytics126📚Education & Research55💼Business & Finance174🏥Healthcare & Wellness22🔍Search & Knowledge20🤖AI Agent Infrastructure208🛡️AI Security & Testing32🧊3D & Spatial22🔎SEO Tools113🏡Real Estate7🗃️Data Extraction104🧠ADHD & Focus Tools11🔬Research & Academia45🧩LLM APIs & Models34⚙️Automation & Workflows45🔐Security & Privacy31📊Analytics & BI55⚖️Legal & Contracts14
Listed in LLM APIs & Models with 37 other toolsPart of 3285+ curated AI tools on AISO
Devstral 2 logo

Devstral 2

Mistral's SOTA open-weight coding model — 72.2% SWE-bench, free API

0
freemiumDR 87Devstral 2 (123B) and Devstral Small 2 (24B) are currently free to use via the Mistral API (console.mistral.ai). Open weights: Devstral 2 ships under a modified MIT license; Devstral Small 2 under Apache 2.0. Self-hosting on compatible hardware is supported. Enterprise pricing available for on-prem deployments.View full pricing →

About Devstral 2

Mistral AI's next-generation open-weight coding model family, released December 9, 2025. Devstral 2 is a 123B-parameter dense transformer with a 256K context window, achieving 72.2% on SWE-bench Verified under a modified MIT license — currently free via the Mistral API. Devstral Small 2 (24B, Apache 2.0) scores 68.0% on SWE-bench Verified and runs on consumer hardware. Up to 7× more cost-efficient than Claude Sonnet at real-world coding tasks per Mistral's human evaluations. Ships alongside Mistral Vibe, an open-source terminal CLI for end-to-end code automation.

Does ChatGPT recommend your AI tool?

If you're building in LLM APIs & Models, run a free AI-visibility scan on your own product — we ask ChatGPT across 5 prompt angles and score how often you get named. ~30 seconds, no signup, no card.

Key Features

72.2% on SWE-bench Verified (Devstral 2, 123B) — state-of-the-art among open-weight models at launch
68.0% on SWE-bench Verified (Devstral Small 2, 24B) — matches models 5× its size
256K context window — supports full codebase ingestion and multi-file edits
Up to 7× more cost-efficient than Claude Sonnet on real-world coding tasks (Mistral human evals)
42.8% win rate vs. 28.6% loss rate against DeepSeek V3.2 in independent human evaluation via Cline
Mistral Vibe CLI: open-source terminal agent for autonomous end-to-end code automation
Multi-file editing: tracks framework dependencies, detects failures, retries with corrections
Fine-tuning support for specific languages or enterprise codebases
Devstral 2: modified MIT license — Devstral Small 2: Apache 2.0
5× smaller than DeepSeek V3.2 (123B vs ~671B) at comparable benchmark performance
Devstral Small 2 runs locally on consumer hardware — single H100 or equivalent
Compatible with Cline, Continue, and other VS Code coding agent integrations

Devstral 2 Pros & Cons

Pros

  • +72.2% on SWE-bench Verified makes it the best open-weight coding model at launch — no closed-source license required
  • +Free API access lowers the bar significantly; most teams can evaluate without a budget discussion
  • +Devstral Small 2 (24B, Apache 2.0) runs on consumer hardware, making local deployment practical
  • +Mistral Vibe CLI provides a ready-made agentic coding interface — no third-party agent framework required
  • +7× cost efficiency vs. Claude Sonnet means the economics work at high token volumes
  • +Permissive licenses (MIT/Apache 2.0) allow commercial use without royalties or usage restrictions

⚠️ Cons

  • Claude Sonnet 4.5 is still significantly preferred in human evals — gap with frontier closed-source models persists
  • Mistral Vibe CLI is new and early; tooling maturity lags behind Copilot and Cursor
  • Benchmark comparisons (SWE-bench %) are self-reported by Mistral; independent reproduction pending at launch
  • Fine-tuning and enterprise on-prem deployment require reaching out to Mistral directly

Who Is Devstral 2 Best For?

👤Developers and teams who need SOTA coding model performance without proprietary-license restrictions
👤Startups running high-volume agentic coding pipelines where cost efficiency vs. Claude Sonnet matters
👤Teams wanting full on-prem control via open weights with no data leaving their infrastructure
👤Engineers building custom coding agents who want to fine-tune on proprietary codebases

Tags

mistralcoding modelopen sourceopen weightscode agentswe-benchagenticllmapisoftware engineeringdevstral
🏷️

Is Devstral 2 your tool?

This is the page buyers and AI assistants read when they look up Devstral 2. Claim your listing for $19 one-time — no subscription, nothing to cancel — and get a Featured badge, top placement in your category, and a permanent dofollow backlink. Prefer it ongoing? Monthly is one click away on the next page.

Stay updated on LLM APIs & Models tools — join our weekly newsletter

One concise email with fresh launches, trending picks, and featured standouts.

Alternatives to Devstral 2

View all Devstral 2 alternatives →

More LLM APIs & Models tools

Agent connectivity: not yet verified