models · zhipu-ai

GLM-4.7-Flash

Zhipu AI GLM-4.7-Flash. 30B compact variant. Arena 759. GPQA Diamond 75.2%, AIME 91.6%. Low-latency deployment of GLM-4 capabilities.

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	4443c1888927bd19c3728a10043a1cd9a2369430efef192ffa70fc22f7be12c2
Method	hf-published-sha256

Re-check this yourself → Source → Full record →

What is recorded

Publisher	zhipu-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	31,221,488,576 (31.2B)
Context window	128K
Modality	text->text
Architecture	glm4_moe_lite
Layers	47
Hidden size	2048
Attention	MLA (20 heads, 20 KV)
Native precision	BF16
Pulls recorded here	290,017
Downloads reported upstream	1,858,331

Reported scores

GPQA Diamond	58.1
IFBench	60.8
Humanity's Last Exam	7.6

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
GeForce RTX 5090	llama.cpp	UD-Q6_K_XL	176 tok/s @ 8,192	candidate	localmaxxing
Radeon AI PRO R9700	llama.cpp	Q2_K_XL	82.0 tok/s @ 2,048	candidate	localmaxxing
Radeon AI PRO R9700	llama.cpp	Q4_K_XL	75.8 tok/s @ 2,048	candidate	localmaxxing
Ryzen AI Max+ 395	llama.cpp	Q4_K_XL	53.0 tok/s @ 4,096	candidate	localmaxxing
Intel Arc Pro B70	llama.cpp	GGUF UD-Q4_K_XL	40.8 tok/s @ 4,096	candidate	localmaxxing
Measured outside 2K–32K context — decode speed depends heavily on context, so these are not comparable with the rows above
GeForce RTX 5090	llama.cpp	Q5_K_M	219 tok/s @ 131,072	candidate	localmaxxing
Radeon AI PRO R9700 ×3 devices	llama.cpp	MXFP4_MOE	97.4 tok/s @ 783	candidate	localmaxxing

From local-ai-registry (MIT), commit 124959d: 13 runs across 6 machines, 0 validated — model revision and runtime pinned, launch accepted — and 13 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.