models · qwen

Qwen3-14B

Qwen3 14B dense instruct model. Efficient mid-range option with strong reasoning and multilingual 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	245877f20829ecc834373c6fc1e86dfb6bc6a699684dbb2967033d4f7a4fc786
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	14,768,307,200 (14.8B)
Context window	128K
Modality	text->text
Architecture	qwen3
Layers	40
Hidden size	5120
Attention	GQA (40 heads, 8 KV)
Native precision	BF16
Pulls recorded here	750,019
Downloads reported upstream	1,584,558

Reported scores

MMLU-Pro	67.5
GPQA Diamond	47
AIME 2025	58
MATH-500	87.1
LiveCodeBench	28
IFBench	23.9
Humanity's Last Exam	4.1

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
RTX PRO 6000 Blackwell	vllm	safetensors FP8	90.8 tok/s @ 32,768	candidate	localmaxxing
GeForce RTX 3080	llama.cpp	Q4_K_M	75.0 tok/s @ 2,048	candidate	localmaxxing
GeForce RTX 5070	llama.cpp	Q4_K_M	66.6 tok/s @ 2,048	candidate	localmaxxing
GeForce RTX 3060	llama.cpp	Q4_K_M	37.0 tok/s @ 2,048	candidate	localmaxxing
GeForce RTX 3060	ollama	Q4_K_M	35.0 tok/s @ 2,048	candidate	localmaxxing
Apple M1 Max 64GB	ollama	Q4_K_M	24.8 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
GeForce RTX 3090	llama.cpp	Q4_K_M	51.7 tok/s @ 1,024	candidate	localmaxxing
GeForce RTX 3090	llama.cpp	Q4_K_M	51.7 tok/s @ 512	candidate	localmaxxing
GeForce RTX 3060 ×2 devices	llama.cpp	Q4_K_M	35.8 tok/s @ 512	candidate	localmaxxing
GeForce RTX 3060	llama.cpp	Q4_K_M	34.6 tok/s @ 512	candidate	localmaxxing

From local-ai-registry (MIT), commit 124959d: 10 runs across 6 machines, 0 validated — model revision and runtime pinned, launch accepted — and 10 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 2026-06-11. 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.