GLM-5.1

Parameters

754B

Published

Apr 7, 2026

GLM-5.1 is an AI model developed by Z.ai (Zhipu AI) (China), first published in April 2026. It works in the language domain, on tasks such as language modeling/generation and question answering.

Epoch AI has no training-compute estimate for this model. The model has 754,000,000,000 parameters.

Epoch AI rates the confidence of this record as confident.

Full record

Organization: Z.ai (Zhipu AI)
Country of organization: China
Domain: Language
Task: Language modeling/generation, Question answering
Parameters: 754,000,000,000
Epoch confidence: Confident

Benchmark results

01	Vending-Bench 2	Score	5634.4
02	WebDev Arena	Arena Score	1534
03	ALE Bench	Performance	887.1
04	Epoch Capabilities Index	ECI Score	150.4
05	OTIS Mock AIME 2024-2025	mean_score	0.92
06	GPQA Diamond	mean_score	0.85
07	SWE-bench Verified	mean_score	0.74
08	WeirdML	Accuracy	0.57
09	SimpleBench	Score (AVG@5)	0.55
10	SciCode	Score	0.44
11	SimpleQA Verified	mean_score	0.37
12	FrontierMath	mean_score	0.33
13	Frontierswe	Dominance	0.32
14	Proofbench	Accuracy	0.22
15	Exploitbench	Mean capability	0.18
16	Chess Puzzles	mean_score	0.18
17	FrontierMath Tier 4	mean_score	0.13
18	Critpt	Accuracy	0.05
19	Gbaeval	Overall score	0