models · deepseek-ai

DeepSeek-R1-0528

Updated DeepSeek-R1 checkpoint (May 2528). Extended reasoning with improved AIME 2025 and GPQA Diamond performance.

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	19d7304244c931dba952ed4ef43517f077c8fd4e7686bfb6732e1ca81f9905d7
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	684,531,386,000 (684.5B)
Context window	128K
Modality	text->text
Architecture	deepseek_v3
Layers	61
Hidden size	7168
Attention	MLA (128 heads, 128 KV)
Native precision	F8_E4M3
Pulls recorded here	820,013
Downloads reported upstream	249,151

Reported scores

MMLU-Pro	84.9
GPQA Diamond	81.3
AIME 2025	76
MATH-500	98.3
LiveCodeBench	77
IFBench	39.6
Humanity's Last Exam	15.8

Reported by a third party and recorded here, not re-run by us. Source: artificialanalysis.ai/api/v2.

Elsewhere in the registry

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

Registry record as of 2026-06-11. 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.