GLM-5.3

Z.ai (Zhipu AI) · zhipu/glm-5-3

Provider	Z.ai (Zhipu AI)
Family	glm
Type	llm-reasoning
Status	active
Released	2026-08-14
Updated	2026-09-04
Parameters	753.3B
Open weights	yes

Lineage

Relationship	Model
Is a continuation of	zai-org/GLM-5.2

Capabilities

Chain Of Thought · tier-2 Function Calling · tier-2 Think Budget Control · tier-2

What it fits on

Ordered by predicted speed. Bandwidth sets decode rate; memory decides whether it runs at all.

Every figure here is computed, not measured. Fit is weights at each quantisation against device memory, with a 25% allowance for the KV cache, activations and the OS. Decode rate is the memory-bandwidth roofline at 70% efficiency. Nobody has run this model on these devices. This is a mixture-of-experts model and no card carries active-parameter counts, so these predictions use total parameters and understate the real speed.

Device	Memory	Bandwidth	Best quality	Weights	tok/s	Fastest
Apple M3 Ultra	512.0 GB	819.0 GB/s	q4	376.66 GB	~1.5	~1.5 q4

Where it runs

Taken from the card's availability section.

Platform	Id on that platform
Fireworks AI	accounts/fireworks/models/glm-5p3
Hugging Face	zai-org/GLM-5.3
Ollama	glm-5.3
OpenRouter	z-ai/glm-5.3
Together AI	zai-org/GLM-5.3
Zhipu GLM	glm-5.3

Verified benchmark evidence

Each row was checked against its source by a reviewer, and carries the model identifier as actually evaluated — which is not always the same as this card's.

Benchmark	Evaluated as	Score	Evidence date	Source kind
aa_briefcase	GLM-5.3 (max)	51.0 normalized Elo percent	2026-09-04 published	independent evaluator	source
aa_lcr	GLM-5.3 (max)	80.0%	2026-09-04 published	independent evaluator	source
automationbench_aa	GLM-5.3 (max)	62.2%	2026-09-07 published	independent evaluator	source
critpt	GLM-5.3 (max)	19.0%	2026-09-04 published	independent evaluator	source
gdp_pdf_aa	GLM-5.3 (max)	12.0%	2026-09-04 published	independent evaluator	source
gdpval_aa	GLM-5.3 (max)	59.0 normalized Elo percent	2026-09-04 published	independent evaluator	source
scicode	GLM-5.3 (max)	59.0%	2026-09-04 published	independent evaluator	source
aa_lcr	GLM-5.3 (max)	79.67%	2026-09-10 evaluated	independent evaluator	source
arena_elo_style_control	glm-5.3-max	1482.04 elo	2026-09-10 evaluated	independent evaluator	source
critpt	GLM-5.3 (max)	19.14%	2026-09-10 evaluated	independent evaluator	source
gdpval_aa	GLM-5.3 (max)	58.75%	2026-09-10 evaluated	independent evaluator	source
gpqa_diamond	GLM-5.3 (max)	91.72%	2026-09-10 evaluated	independent evaluator	source
scicode	GLM-5.3 (max)	59.03%	2026-09-10 evaluated	independent evaluator	source

Reported benchmark scores

This card reports no benchmark scores yet.

Benchmark	Catalogue standing	Score

Data

This card as JSON · See it in the graph · Edit on GitHub