MODEL / MISTRAL AI

Devstral Small 2 24B

Model card checked 2026-09-16 · Memory fit estimated; speed unknown unless a separately linked measurement matches your configuration.

Also searched as: devstral-small-2

The 256K maximum context is not a practical default on consumer memory.

Primärquelle

Gesamtparameter24B

Geschätzter Speicher~18.5 GBGGUF Q4 / BF16

Kontext256KText + Image

LicenseApache-2.0Multilingual

01 / GEEIGNET FÜR

coding agents
repository work
tool use

02 / GESCHÄTZTE PASSFORM

Mac Studio M5 Ultra · 96 GBBequem Mac mini M6 · 24 GBKnapp MacBook Pro M5 Max · 64 GBBequem Framework Desktop · Ryzen AI Max+ 395 · 128 GBBequem Radeon AI PRO R9700 · 32 GBBequem

03 / STARTEN MIT

These are runtime examples, not validated installation recipes. Fit estimates assume quantized files; official repositories may hold full-precision weights. MLX, GGUF and TensorRT require matching conversions or engines.

OllamaEasyollama run <model-tag>

llama.cppMediumllama-server -m /path/to/verified-model.gguf --port 9931

LM StudioEasySearch “mistralai/Devstral-Small-2-24B-Instruct-2512” → Load a Q4/INT4 build → Start Server

MLX LMMediummlx_lm.server --model mistralai/Devstral-Small-2-24B-Instruct-2512

Achtung

The 256K maximum context is not a practical default on consumer memory.