models · Qwen

Qwen3-Coder-30B-A3B-Instruct

Catalogued from llm-stats Open LLM Leaderboard at Quarantine — unverified, pending attestation.

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.

Method	hf-published-sha256

Re-check this yourself → Source → Full record →

What is recorded

Publisher	Qwen
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	Apache-2.0
Parameters	30,532,122,624 (30.5B)
Context window	262,144 tokens
Modality	text->text
Architecture	qwen3_moe
Layers	48
Hidden size	2048
Attention	GQA (32 heads, 4 KV)
Native precision	BF16
Pulls recorded here	6
Downloads reported upstream	604,411

Reported scores

MMLU-Pro	70.6
GPQA Diamond	51.6
AIME 2025	29
MATH-500	89.3
LiveCodeBench	40.3
IFBench	32.7
Humanity's Last Exam	3.8

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

Measured on real hardware

None of these runs are ours. Each figure is the reporter’s, linked to their own evidence, under their own trust label. The number shown is single-stream decode — one user, concurrency 1 — because aggregate throughput across many users is a different quantity and reads as far faster than anyone will see.

Reported to local-ai-registry

Hardware	Engine	Build measured	Decode, 1 stream	Label	Reported by
Intel Arc Pro B70	llama.cpp	GGUF UD-Q4_K_XL	108 tok/s @ 4,096	candidate	localmaxxing
Apple M3 Ultra 512GB	llama.cpp	Q4_K_M	100 tok/s @ 8,192	candidate	localmaxxing
Intel Arc Pro B70 ×2 devices	llama.cpp	UD-Q4_K_XL	93.3 tok/s @ 2,048	candidate	localmaxxing
Ryzen AI Max+ 395	llama.cpp	Q5_K_M	80.5 tok/s @ 16,384	candidate	localmaxxing
Apple M1 Max 64GB	ollama	Q4_K_M	72.3 tok/s @ 2,048	candidate	localmaxxing
Ryzen AI Max+ 395	llama.cpp	Q4_K_M	70.5 tok/s @ 4,096	candidate	localmaxxing
Radeon RX 7900 XTX	llama.cpp	Q4_K_M	43.8 tok/s @ 2,048	candidate	localmaxxing
Measured outside 2K–32K context — decode speed depends heavily on context, so these are not comparable with the rows above
Radeon PRO W7800	llama.cpp	Q4_K_M	109 tok/s @ 512	candidate	0xsero
Radeon RX 7900 XTX	llama.cpp	Q4_K_M	46.9 tok/s @ 1,024	candidate	localmaxxing
Radeon PRO W7800 ×2 devices	llama.cpp	Q4_K_M	35.6 tok/s @ 512	candidate	0xsero

From local-ai-registry (MIT), commit 124959d: 14 runs across 9 machines, 0 validated — model revision and runtime pinned, launch accepted — and 14 candidate, their label for useful evidence without a reproducible-launch promise.

Elsewhere in the registry

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

Registry record as of 2026-09-26. 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.