Skip to content

Mistral AI

Mistral Large 4 (26.10)

1,050 B params (52 B active) · 1M context · Open weights

Released 2026-11-05

Details

Provider
Mistral AI
Family
Mistral Large
Context length
1M tokenssource
Modalities
Text, imagesource
Open weights
Yessource
Licence
Mistral Ecosystem Open Weight Licensesource
Parameters (total)
1050 Bsource
Active parameters (per token)
52 Bsource

API pricing

Input$0.7 per million tokenssource

Output$2.1 per million tokenssource

Checked on 2026-11-12

Running it locally

Memory needed for 1050 B parameters (mixture of experts: 52 B active per token, but all 1050 B must be held in memory) 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_M641.8 GB1.0 GB32.6 GB675.4 GB
Q6 K861.3 GB1.0 GB43.6 GB905.9 GB
Q8 01115.6 GB1.0 GB56.3 GB1172.9 GB
FP162100.0 GB1.0 GB105.5 GB2206.5 GB

Catalogued hardware that fits at Q4

No catalogued hardware has enough memory yet, or none is listed.

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.