Title: Devstral Small 2 (24B) %E2%80%94 devstral-small-2-24b Canonical: https://llmeu.com/nl/models/devstral-small-2-24b/ Markdown mirror: https://llmeu.com/nl/models/devstral-small-2-24b.md Machine-readable: https://llmeu.com/llms.txt (index) - https://llmeu.com/llms-full.txt (full catalogue) - https://llmeu.com/llms-catalogue.json Provider: Mistral AI Licence: apache-2.0 Hosting: catalogue only %E2%80%94 not hosted on LLM EU yet Sources: https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512 # Devstral Small 2 (24B) [Mistral AI](/nl/providers/mistral/) %C2%B7 Providers EU Laatst beoordeeld: 14 okt 2026 ## Soevereiniteit en hosting No endpoint for this model is reachable right now (de-ber Gepland onderhoud). The region and prices describe the deployment, not current availability. ## Samenvatting Devstral Small 2 24B (devstral-small-2-24b) is Mistral's Apache-2.0 coderingsmodel, afgestemd met All Hands AI voor agents die bestanden bewerken en tools aanroepen. Alleen tekst, een contextvenster van 256k tokens, en een modelkaart die 68.0% rapporteert op SWE-bench Verified en 55.7% op SWE-bench Multilingual. ## Specificaties - Provider Mistral AI (FR) - Contextvenster 262k - Parameters 24.01B - Modaliteiten text, code - Licentie apache-2.0 %C2%B7 [licence](https://www.apache.org/licenses/LICENSE-2.0) - Open weights yes - Releasedatum 28 nov 2025 - Taken [Coderen](/nl/tasks/coding/), [Retrieval (RAG)](/nl/tasks/rag/), [Redeneren](/nl/tasks/reasoning/) ## Talen **Sterk:** en **Voldoende:** fr, de, es, it, pt, zh, ja, ko #### Sterk - en #### Voldoende - fr - de - es - it - pt - zh - ja - ko Niet geverifieerd: de provider heeft deze claim niet gedocumenteerd. %E2%80%94 a language is listed as strong only where a source claims it; the card links the evidence where one exists. ## Wanneer te gebruiken #### Sterke punten - agentic coding: multi-file edits and tool use in repositories - long 256k-token context window for large codebases - Apache-2.0 licence permits commercial self-hosting - vendor-published SWE-bench Multilingual score #### Beperkingen - text-only, no image input - benchmark scores are self-reported by the vendor - no evaluation has been run by LLM EU ## Veelgestelde vragen ### What is Devstral Small 2 (24B) built for? Software-engineering agents: the card describes a model fine-tuned for that with All Hands AI, for multi-file edits and tool use in repositories, with a long context window for large codebases. It is text-only %E2%80%94 the vision encoder of its Mistral Small base was removed during fine-tuning, and the card lists "text-only, no image input" as a limit. Source: https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512. ### How many parameters does Devstral Small 2 (24B) have, and what is its context window? The card records 24.0B (safetensors total 24,011,361,840) (source: https://huggingface.co/api/models/mistralai/Devstral-Small-2-24B-Instruct-2512) over a 262144-token context window (source: https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512). ### What scores does Devstral Small 2 (24B) report? SWE-bench Verified 68.0% (vendor-reported), SWE-bench Multilingual 55.7% (vendor-reported) and Terminal Bench 2 22.5% (vendor-reported), all vendor-reported. Source: https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512. ### Can Devstral Small 2 (24B) be self-hosted commercially? Yes, under Apache License 2.0 (https://www.apache.org/licenses/LICENSE-2.0): the card lists "Apache-2.0 licence permits commercial self-hosting" among its strengths, and the weights are downloadable. The licence text governs. ### Is Devstral Small 2 (24B) served on LLM EU? It has no endpoint in the catalogue, so it is catalogue only %E2%80%94 not hosted on LLM EU yet, with no region and no retention to report. A card whose status is "listed" is a catalogue entry, not a running deployment. ### Prijzen Dit model is opgenomen in de catalogus ter referentie. Het wordt nog niet gehost op LLM EU en er is hier geen first-party endpoint dat het aanbiedt. Verzoeken eraan gaan rechtstreeks naar de provider, onder de eigen voorwaarden van die provider. ### Verificatiestatus Waar een veld geen bron heeft, toont de kaart het als ongeverifieerd in plaats van het in te vullen. Een leeg veld is informatie. LLM EU Verified betekent gedocumenteerd, niet wettelijk gecertificeerd. Het is geen conformiteitsbeoordeling onder de AI-verordening en LLM EU is geen EU-orgaan. ### Versies en varianten - 2025-11-28-q8 %C2%B7 q8 %C2%B7 [weights](https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512) - catalogue %C2%B7 Alleen catalogus %C2%B7 [weights](https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512) ### Voorbeeldverzoek ``` curl https://api.llmeu.com/v1/chat/completions \ -H "Authorization: Bearer $LLMEU_API_KEY" \ -H "Content-Type: application/json" \ -d '{"model":"devstral-small-2-24b","messages":[{"role":"user","content":"Hello"}]}' ``` Er is geen playground uitgerold voor dit model: geen enkel endpoint bedient het, dus een verzoek zou nergens heen kunnen. ### Benchmarks - SWE-bench Verified: **68.0% (vendor-reported)** [source](https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512) - SWE-bench Multilingual: **55.7% (vendor-reported)** [source](https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512) - Terminal Bench 2: **22.5% (vendor-reported)** [source](https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512) ### Bronnen - [huggingface.co](https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512) ### Changelog van de kaart - 28 nov 2025 %E2%80%94 Releasedatum - 14 okt 2026 %E2%80%94 Laatst beoordeeld Laatst beoordeeld: 14 okt 2026 ### Taken - [Coderen](/nl/tasks/coding/)