Skip to content

DeepSeek

DeepSeek-V4-Pro-0813

1,650.5 B params · 1M context · Open weights

Released 2026-08-13

Details

Provider
DeepSeek
Family
DeepSeek V4
Context length
1M tokenssource
Open weights
Yessource
Licence
MITsource
Parameters (total)
1650.5 Bsource
Api id
deepseek-v4-prosource
Max output tokens
384000 tokenssource

API pricing

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

Running it locally

Memory needed for 1650.498 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_M1008.9 GB1.0 GB50.9 GB1060.8 GB
Q6 K1353.9 GB1.0 GB68.2 GB1423.1 GB
Q8 01753.7 GB1.0 GB88.2 GB1842.8 GB
FP163301.0 GB1.0 GB165.5 GB3467.5 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.