DeepSeek-V3

A 671B-total, 37B-active MoE model that extended the efficient-training and open-release path.

LegacyLanguageCodingOpen weights

MODEL NUMBER017

Specifications

Only published specifications are shown.

Total parameters

671B

Active parameters

37B

Architecture

MoE

Context

128K

Input modalities

Text

Open weights

Yes

License

MIT

Status

Legacy

A 671B-total, 37B-active MoE model that extended the efficient-training and open-release path.

V3 pushed open MoE general and coding capabilities toward the frontier and became the base for subsequent DeepSeek reasoning work.

PreviousDeepSeek-V2236B · 128K ctx

CurrentDeepSeek-V3671B · 128K ctx

New capabilities Tool use Multilingual

MLA

CodingTool useMultilingual

2024-12-26Open weights

DeepSeek-V3 launched as an open 671B MoE.

Select a card to view the related model.