Skip to content

Mistral AI

Mistral Small 3.2 24B Instruct

24 B params · 131072 context · Open weights

Released 2025-06-19

Details

Provider
Mistral AI
Family
Mistral Small
Context length
131072 tokenssource
Open weights
Yessource
Licence
Apache 2.0source
Parameters (total)
24.01 Bsource

API pricing

No hosted API price verified. Open-weight models can be self-hosted.

Running it locally

Memory needed for 24.011 B parameters at 8,192 tokens of context. Architecture not recorded, so a flat 1.0 GB KV allowance is used. Estimates, not measurements.

FormatWeightsKV cacheOverheadTotal
Q4 K_M14.7 GB1.0 GB1.2 GB16.9 GB
Q6 K19.7 GB1.0 GB1.5 GB22.2 GB
Q8 025.5 GB1.0 GB1.8 GB28.3 GB
FP1648.0 GB1.0 GB2.9 GB51.9 GB

Catalogued hardware that fits at Q4

Mini PCs reserve 8.0 GB for the operating system. “Comfortable” keeps at least 10% memory spare.

Benchmarks

No benchmarks for this model yet

We only publish results we can evidence, so there is nothing here yet and we have not filled the gap with estimates. Here is how we will test and label results.