whatmodel.sh / Devstral 2 (2512)

Devstral 2 (2512)

+ contribute

Devstral 2 (2512) is an open-weights language model from Mistral, released December 2025. Its published size is 123B parameters and a 256k-token context window. Licence: modified-mit.

Catalogue identity notes

Arena entry 'devstral-2' (webdev). Maker repo mistralai/Devstral-2-123B-Instruct-2512 @1613bf0: licence tag other. Parameters: card table: Devstral 2 'Size (B Parameters) 123' (dense). Released 2025-12: Mistral model card 'Devstral 2' dated December 9, 2025. Mistral card status 'Deprecated' (5/22/2026); retirement date not confirmed, so no current API is asserted. Open weights (Modified MIT). Distinct from the catalogued devstral-small-2.

62.2±1.7#176 of 719 LLMsfrom 5 benchmark results across 3 sources

$ how it compares

The LLMs ranked either side of Devstral 2 (2512), out of 828 with a score, then more from Mistral.

#	model	score	where	RAM	released
204	Nemotron 3.5 Lightning 30B-A3B NVIDIA	62±2	api $0.087/M +local	~20GB	2026-08
205	GLM-4.6V Z.ai	62±2	api $0.45/M +local	~51GB	2025-12
206	Devstral 2 (2512) Mistral	62±2	local	~75GB	2025-12
207	Gemini 2.5 Flash Preview (05-20) Google	62±1	comparison only		2025-05
208	A.X K1 (519B-A33B) SK Telecom	62±1	local	~1091GB	2025-12
	more from Mistral
79	Mistral Large 4 (public preview) Mistral	71±2	api $2.06/M		2026-10
174	Mistral Medium 3.5 (128B) Mistral	65±1	api $3.00/M +local	~42GB	2026-04
180	Mistral Medium 3.1 (2508) Mistral	64±2	comparison only		2025-08

see all LLMs · find the ones that fit your machine

$ what it's good at

agent tasks█··········█ #1 of 2 · 37% of its score · 3 tests

test	result	vs the others in its run	difficulty	counts	why
SWE-bench Verified lab	72.2% ✎	▁▁▁▁▁▁▂▄█▆▁▁ beats 85/131	hard · avg 32%	▇▇▇▇▇▇ 15%	sharp · at its level · small run
Terminal-Bench 2.0 lab	32.6% ✎	▂▂▅▅████▆▂▃▃ beats 7/36	very hard · avg 11%	▇▇▇▇▇ 13%	at its level · small run
SWE-bench Multilingual lab	61.3% ✎	▁··▁▃▂▅▇█▃▃▁ beats 21/60	hard · avg 30%	▇▇▇▇ 10%	at its level · small run

coding▁▁▁▃▆██▇▆▄▃▁ #179 of 458 · 69% of its score · 3 tests

Beats Qwen3.5 9B here, though they rank higher overall. Trails QwQ-32B-Preview, OpenThinker 32B, Ministral 3 8B Reasoning 2512 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Arena WebDev (Code Arena)	1196.1 ✎	▁▃▄▃▅█▄▇▆▄▁▁ beats 5/119	avg 981	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 44%	sharp
SWE-bench Verified lab	72.2% ✎	▁▁▁▁▁▁▂▄█▆▁▁ beats 85/131	hard · avg 32%	▇▇▇▇▇▇ 15%	sharp · at its level · small run
SWE-bench Multilingual lab	61.3% ✎	▁··▁▃▂▅▇█▃▃▁ beats 21/60	hard · avg 30%	▇▇▇▇ 10%	at its level · small run

other tests1 not in a task · 17% of its score

test	result	vs the others in its run	difficulty	counts	why
DSEffi-Bench Pass (1,000 data-science tasks, greedy)	43.0% ✎	▅··▃▅·█·▃██▃ beats 7/15	hard · avg 24%	▇▇▇▇▇▇▇ 17%	at its level

$ where to run it

llama.cpp · recommended

version

Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
Windows
irm https://llama.app/install.ps1 | iex
other ways to install ↗
Run Devstral 2 (2512) · needs ~74.9 GB · from huggingface.co/lmstudio-community
chat in the terminal
llama cli -hf lmstudio-community/Devstral-2-123B-Instruct-2512-GGUF:Q4_K_M
Or start a local server and open the address it prints (http://localhost:9931)
llama serve -hf lmstudio-community/Devstral-2-123B-Instruct-2512-GGUF:Q4_K_M

LM Studio

version

Install LM Studio · Open the app once before using the lms command.
download it (macOS / Windows / Linux) ↗
Run Devstral 2 (2512) · needs ~74.9 GB · from huggingface.co/lmstudio-community
open in LM Studio ↗
Or download it in the terminal
lms get https://huggingface.co/lmstudio-community/Devstral-2-123B-Instruct-2512-GGUF@Q4_K_M -y
Then run lms ls to see its name, and lms chat followed by that name

Ollama

version

Install Ollama
download it (macOS / Windows) ↗
Linux
curl -fsSL https://ollama.com/install.sh | sh
Run Devstral 2 (2512) · needs ~74.3 GB · from ollama.com
chat in the terminal
ollama run devstral-2:123b-instruct-2512-q4_K_M

MLX (Mac)

version

Install MLX (Mac) · Apple silicon Macs only, text only.
with uv
uv tool install mlx-lm
or
pip install mlx-lm
Run Devstral 2 (2512) · needs ~70.3 GB, Mac only · from huggingface.co/mlx-community
chat in the terminal
mlx_lm.chat --model mlx-community/Devstral-2-123B-Instruct-2512-4bit
Or start an OpenAI-compatible server at http://localhost:8080/v1
mlx_lm.server --model mlx-community/Devstral-2-123B-Instruct-2512-4bit

Weights and documentation: huggingface.co/mistralai ↗

$ how the score was worked out · LMArena leaderboard dataset 44%, mistralai/Devstral-Small-2-24… 37%, DSEffi-Bench: Demystifying La… 17%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

source	results	share of the score	licence
LMArena leaderboard dataset	1	▇▇▇▇▇▇▇▇▇▇▇▇▇ 44%	CC BY 4.0
mistralai/Devstral-Small-2-24B-Instruct-2512 model card · lab-reported	3	▇▇▇▇▇▇▇▇▇▇▇ 37%	Apache-2.0
DSEffi-Bench: Demystifying Large Language Models' Capability in Efficient Data Science Code Generation — authors' own model runs	1	▇▇▇▇▇ 17%	CC BY 4.0
prior · pull toward the middle		▇ 1%
weighted average		= 62.2

$ improve this page

Update model information Add a place to run it Add a test result Propose a benchmark Correct a benchmark score Edit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated 2026-10-11 02:20 UTC · sources and licences · json