Devstral Small 2 24B Instruct 2512
Hugging Face (mistralai) Open weights
The catch: The card says it fits a single RTX 4090 or a 32GB Mac, but its own vLLM example uses two GPUs to serve the full 256k context.
| Access | Open weights |
|---|---|
| Free limits | Free to download and self-host; you pay only for your own compute. Apache 2.0, so no user cap. 24B parameters (FP8 weights). |
| Modality | text, code, image |
| Credit card | Not required |
| Commercial use | Allowed |
| Trains on your data | – |
| Licence or terms | Apache-2.0 |
| Context window | 262,144 tokens |
| Model id | mistralai/Devstral-Small-2-24B-Instruct-2512 |
| Base URL | – |
| Last checked | 2 Oct 2026 |
How to use Devstral Small 2 24B Instruct 2512
Download the weights and serve them yourself; the quickstart below shows one way. Licence: Apache-2.0.
Quickstart
ollama run devstral-small-2
Paste into Claude Code
I want to run the open-weight model "Devstral Small 2 24B Instruct 2512" (mistralai/Devstral-Small-2-24B-Instruct-2512) for my project. Please: 1. Check the model card and tell me the hardware it needs and the simplest way to serve it (vLLM, SGLang, Ollama or a hosted endpoint). 2. Give me a minimal working setup and a test request. 3. Tell me the licence terms for commercial use. 4. If a cheaper or smaller variant would do the job, mention it.
Before you ship on it
- The card says it fits a single RTX 4090 or a 32GB Mac, but its own vLLM example uses two GPUs to serve the full 256k context.
- Apache 2.0; the HF model card says it allows usage and modification for both commercial and non-commercial purposes.
- Free to run means you pay for the GPUs, the serving stack and the on-call.
Questions
Is Devstral Small 2 24B Instruct 2512 free?
Yes. Open weights offers it as open weights, with these limits: Self-host free · Apache-2.0. No credit card is needed.
Can I use Devstral Small 2 24B Instruct 2512 commercially?
Yes, under Apache-2.0. Check the current terms before you ship.
How do I start with Devstral Small 2 24B Instruct 2512?
Download the weights and serve them yourself; the quickstart below shows one way. Licence: Apache-2.0.
Source: huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512, checked 2 Oct 2026.