Devstral 2 2512
2123°
Run this model
Offers / Ads
Mistral | 262,144 token context | $0.40/M input, $2.00/M output. Devstral 2 is a state-of-the-art open-source model by Mistral AI specializing in agentic coding. It is a 123B-parameter dense transformer model supporting a 256K context window. Devstral 2 supports exploring codebases and orchestrating changes across multiple files while maintaining architecture-level context. It tracks framework dependencies, detects failures, and retries with corrections—solving challenges like bug fixing and modernizing legacy systems. The model can be fine-tuned to prioritize specific languages or optimize for large enterprise codebases. It is available under a modified MIT license.
Guides for this model
Best local LLM by use case (October 2026): the right open model for every job
How much VRAM do I actually need for local AI?
Open weights vs open source vs API-only: what you are actually allowed to do
What is a GGUF file? Quantized models explained (Q4, Q8, fp8, NF4)
Captain's notes
Notes from 2026-10-13
Devstral 2 is Mistral AI's 123B-parameter open-weights model built specifically for agentic software engineering. It navigates large codebases, edits across multiple files simultaneously, and tracks framework dependencies — all within a 256K context window. If you are building or running an AI coding agent rather than just asking for snippets, this is the model to reach for first among open-weights options.
Known for
- Agentic codebase work: explores, edits and orchestrates changes across multiple files.
- 72.2% on SWE-bench Verified — competitive with models several times its parameter count.
- 61.3% on SWE-bench Multilingual, strongest multilingual SWE score among open-weights peers listed.
- 256K context window holds architecture-level understanding of large codebases.
- Works with Cline, OpenHands, SWE-Agent, Claude Code and Mistral's own Vibe CLI.
Watch out for
- At 123B parameters, self-hosting demands serious hardware; check your VRAM before downloading.
- Modified MIT licence — not a standard MIT; read the terms before commercial or derivative use.
- Closed frontier models (GPT 5.1 Codex Max, Gemini 3 Pro, Claude Sonnet 4.5) still outscore it on SWE-bench.
Prompts to try
Multi-file refactor plan
I have a Python Flask monolith where authentication logic is scattered across app.py, routes/user.py and utils/helpers.py. Analyse the dependencies between these files, identify every place auth state is read or written, and produce a step-by-step plan to extract it into a standalone auth module without breaking existing behaviour.
Legacy modernisation audit
Here is a Node.js 12 codebase that uses callbacks throughout and depends on several deprecated npm packages. List every pattern that needs to change to reach Node.js 20 compatibility, group them by risk level, and suggest the order in which to tackle them so each change is independently testable.
Bug hunt with retry logic
A Django REST API returns a 500 error intermittently when two users update the same object within milliseconds of each other. Walk through the likely race conditions in a typical Django ORM setup, write a test that reliably reproduces the issue, then propose and implement a fix using database-level locking.
Compare with
- Mistral Large 4API-only model1,048,576 token context · $0.68/M input, $2.09/M output
- Mistral Medium 3.5API-only model262,144 token context · $1.50/M input, $7.50/M output
- Mistral Large 2407API-only model131,072 token context · $2.00/M input, $6.00/M output
- Mistral NemoOpen-weights model131,072 token context · $0.03/M input, $0.03/M output
More text models from mistralai
Questions people ask
- Are the weights open and can I run Devstral 2 locally?
- Yes — the weights are available on Hugging Face under a modified MIT licence, so you can download and self-host the model, though 123B parameters at FP8 precision requires substantial GPU memory.
- What does Devstral 2 cost via the API, and what is its context limit?
- Mistral charges $0.40 per million input tokens and $2.00 per million output tokens, with a 256K (262,144-token) context window.
- How does Devstral 2 compare to Devstral Small 2?
- Devstral 2 at 123B scores 72.2% on SWE-bench Verified versus 68.0% for the 24B Devstral Small 2; the larger model also leads on multilingual SWE tasks and TerminalBench 2, at the cost of higher hardware and API spend.
Captain's verdictThe strongest open-weights coding agent available at this size — if you can feed it enough VRAM.
Model card on Hugging Face
Captain writes these notes from the model's own description and model card. Prices and context sizes are the provider's at the time of writing and can change.
Ranking badge
Updated Oct 22, 2026
Made this model? Show this badge on your page.
Paste it into your Hugging Face card or Civitai description.
How this score adds up
This model has no popularity data from any source yet, so it carries a stand-in number below 7,000° that keeps the lists complete. It never ranks above a model with data. The breakdown appears here as soon as a source reports it.
Glossary: the terms used here
- Heat score
- One number for how much a model is used and wanted right now, from 0° to 9,999°. Every list on this site that says hottest is sorted by it and nothing else.
- Floor
- The 7,000° every model with real data starts from. The scale runs from 0° to 9,999°; the part below the floor is reserved for stand-ins (models without data), so that a model with data always ranks above one without. The floor is the same for everyone and decides nothing; the degrees above it are the measurement.
- Download-equivalents
- Everything the score treats as a download, added up: downloads themselves, each torrent seeder as 1,000, each AI Horde image and each run on a generation site as one.
- Log scale
- Counting by tens, not by ones: going from 1,000 to 10,000 downloads adds the same degrees as going from 10,000 to 100,000. It keeps giants from crowding out everything else.
- Likes
- Thumbs up and likes from every source, worth half a download each on the log scale.
- Adoption rate
- How many downloads a day a model gets right now, from the difference between two readings of a source. A quarter of the weight of the download term.
- Measured and estimated
- Measured means two readings of a source exist and the difference is real. Estimated means only today's total is known and the past is spread over the model's life; the site says which one it is showing.
- Seeder
- Someone who keeps a copy of the weights online for others to download over a torrent, or a hosted mirror doing the same. Each one counts as 1,000 downloads.
- AI Horde
- A volunteer network that generates images for free on donated computers. The number of images it has made with a model is public, and each counts as one download.
- Generation sites
- Sites that host their own copies of a model and let people generate with it, such as PixAI, SeaArt and Yodayo. Their downloads count like Civitai's and each generation run as one download.
- OpenRouter tokens
- A text model served through an API, such as Claude or GPT, has no weights to download, so its usage is measured in tokens: the pieces of text it read and wrote through OpenRouter, a gateway many apps use, over the last 30 days. Every 10,000,000 tokens count as one download. Source: OpenRouter (openrouter.ai/rankings), CC BY 4.0.
- Full model (checkpoint)
- The weights themselves, the file that generates on its own. It gets a bonus because add-ons depend on it.
- LoRA and embedding
- Small add-on files that steer someone else's weights toward a style, a character or a concept. They cannot run alone.
- Family and weights
- The base a model is built on, such as SD 1.5, SDXL, Flux or Wan. Every model inherits the release date of its family's weights.
- Age
- Degrees taken off for how long ago the family's weights came out: 150° a year, up to 750°. It keeps old downloads that had years to pile up from freezing the top of the list.
- Family boost
- A small, bounded nudge for LoRAs of a family this site is pushing forward right now, always listed in the breakdown.
- Stand-in score
- A stable number below the floor given to a model no source reports on yet, so lists stay complete. It never ranks above a model with data.
- New! (grace period)
- A model less than 7 days old shows New! instead of a score and holds no place in the rankings. Its score is computed from day one, but a model one day old grows faster than anything, so without the wait the week's #1 would always be whatever came out yesterday rather than what held up for a week.
- Trend
- How a model's growth over a period compares with the typical model's growth over the same period. Faster than typical reads as a rise, slower as a fall.
- Typical model
- The middle of the catalog: half the models with a figure grew faster over the period, half slower.
Devstral 2 2512 is a text model for the Text / LLM family listed on ModelTrends.ai, a read-only catalog of open source AI image, video and text models. Compare it with other Text / LLM models and check its heat score to see how it is trending.
Browse by Model Family
Text / LLM Models
434
Anima Models
528
Chroma Models
19
Fibo Models
2
Flux Models
1,496
Flux 2 / Klein Models
1,239
Flux 3 Models
3
Moss Models
3
Iris Models
1
MiniMax H3 Models
364
Hunyuan Models
204
Ideogram Models
20
Krea2 Models
717
Qwen Models
20
Qwen2 Models
47
Ltx2 Models
153