models · deepseek-ai

DeepSeek-V3.2 (Non-thinking)

DeepSeek V3.2 non-thinking mode. 685B MoE, 131K context, 318 tok/s. Arena 980. High-throughput dense-equivalent inference without chain-of-thought overhead.

Compare Verify

What this registry has checked

Quarantine. We hold the publisher’s own published hash for these weights and have recorded it against this entry. We have not downloaded and hashed the weights ourselves, so the hash is the publisher’s claim, not our measurement.

Weight file SHA-256	fd2fe54fb74c820540b4fc21f59b49123d5f6030a24cebbeffd36c92ad253838
Method	hf-published-sha256

Re-check this yourself → Source → Full record →

What is recorded

Publisher	deepseek-ai
Origin	Recorded as a covered nation under 10 U.S.C. § 4872(f) — the PRC, Russia, Iran or North Korea. The registry did not record which.
Licence	MIT
Parameters	685,396,921,376 (685.4B)
Context window	131K
Modality	text->text
Architecture	deepseek_v3
Layers	61
Hidden size	7168
Attention	MLA (128 heads, 128 KV)
Native precision	F8_E4M3
Pulls recorded here	870,035
Downloads reported upstream	1,959,027

Elsewhere in the registry

All models → Browse with filters → What will run on my hardware? →

Registry record as of 2026-09-15. Identity, licence and parameter count are properties of a release and do not move; tier can, and the verify link above re-checks it against the live registry rather than this page.