What are the hardware requirements for Devstral Small 2 24B Instruct 2512?

Devstral Small 2 24B Instruct 2512 requires approximately 51.6 GB of VRAM and supports a context window of 384K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Devstral Small 2 24B Instruct 2512 and how large is it?

Devstral Small 2 24B Instruct 2512 is developed by mistralai, a model with 24b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Devstral Small 2 24B Instruct 2512?

Devstral Small 2 24B Instruct 2512 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Devstral Small 2 24B Instruct 2512 by mistralai — VRAM 51.6GB, 384K context

Name: Devstral Small 2 24B Instruct 2512
Author: mistralai

Devstral Small 2 24B Instruct 2512 is an open-source language model by mistralai. Features: 24b LLM, VRAM: 51.6GB, Context: 384K, License: MNPL-0.1, Instruction-Based, LLM Explorer Score: 0.35.

Arxiv:2501.19399 Base model:mistralai/mistral-s... Base model:quantized:mistralai... Deploy:azure Fp8 Instruct Mistral-common Mistral3 Region:us Safetensors Sharded Tensorflow Vllm

Model Card on HF 🤗: https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512

Devstral Small 2 24B Instruct 2512 Parameters and Internals

LLM Name	Devstral Small 2 24B Instruct 2512
Repository 🤗	https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512
Base Model(s)	mistralai/Mistral-Small-3.1-24B-Base-2503 mistralai/Mistral-Small-3.1-24B-Base-2503
Model Size	24b
Required VRAM	51.6 GB
Updated	2026-06-19
Maintainer	mistralai
Instruction-Based	Yes
Model Files	19.8 GB: 1-of-2 6.0 GB: 2-of-2 4.9 GB: 1-of-6 5.0 GB: 2-of-6 5.0 GB: 3-of-6 5.0 GB: 4-of-6 5.0 GB: 5-of-6 0.9 GB: 6-of-6
Model Architecture	MambaSSM
License	MNPL-0.1
Context Length	393216
Model Max Length	393216
Tokenizer Class	TokenizersBackend
Padding Token	<pad>
Vocabulary Size	131072

Rank the Devstral Small 2 24B Instruct 2512 capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Quantized Models of the Devstral Small 2 24B Instruct 2512

Model	Likes	Downloads	VRAM
...Small 2 24B Instruct 2512 4bit	8	104813	15 GB
...l 2 24B Instruct 2512 MLX 8Bit	6	3915	25 GB
...Small 2 24B Instruct 2512 6bit	5	531	20 GB
...Small 2 24B Instruct 2512 8bit	1	422	26 GB

Best Alternatives to Devstral Small 2 24B Instruct 2512

Best Alternatives	Context / RAM	Downloads	Likes
...Small 2 24B Instruct 2512 Bf16	384K / 47.8 GB	367	4
...al Small 3.2 24B Instruct 2506	128K / 48.2 GB	312176	602
Devstral Small 2507	128K / 47.3 GB	116445	368
Magistral Small 2509	128K / 48.2 GB	33508	305
Devstral Small 2505	128K / 47.3 GB	4960	867
...mall 3.2 24B Instruct 2506 FP8	128K / 25.8 GB	81	6
Magistral Small 2506	40K / 47.3 GB	18796	609
Magistral Small 2507	40K / 47.3 GB	632	104
...al Small 3.1 24B Instruct 2503	0K / 48.2 GB	298064	1378
...istral Small 24B Instruct 2501	0K / 47.3 GB	85437	962

Note: green Score (e.g. "73.2") means that the model is better than mistralai/Devstral-Small-2-24B-Instruct-2512.