Back to models

GLM-5.3

ECI#24
155
Z.ai (Zhipu AI)
Aug. 14, 2026
Closed weights

GLM-5.3 is a closed-weights model from Z.ai (Zhipu AI), released August 14, 2026. It has an ECI score of 155 and ranks 24th of 247 tracked models. At the time of its release, it ranked 22nd of 244 tracked models. At $1.40/Mtok for input tokens and $4.40/Mtok for output tokens, it is less expensive per token than similarly-capable models.

Input price$1.40 / 1M tokens
Output price$4.40 / 1M tokens
Parameters7.4e11
Training computeUnknown

GLM-5.3 vs select alternatives

DomainBenchmark
GLM-5.3
Z.ai (Zhipu AI)
Aug. 14, 2026
Closed weights
Z.ai (Zhipu AI)
Jun. 16, 2026
Open weights
DeepSeek
Aug. 13, 2026
Open weights
Google DeepMind
Jul. 21, 2026
Closed weights
ECI scoreBest score 167
155152155154
Ranking#24#37#21#29
Agent
Vending-Bench 2Best score 11181.9
8163.68313.8
Games
Chess PuzzlesBest score 72.0%
21.0%
21.0%
47.0%
43.0%
Mystery Game PuzzlesBest score 84.0%
33.0%
19.0%
43.0%
30.0%
Mathematics
68.8%
59.2%
64.6%
58.9%
FrontierMath Tier 4 (v2)Best score 97.6%
29.3%
29.3%
26.8%
22.0%
OTIS Mock AIME 2024-2025Best score 100.0%
91.1%
86.4%
98.6%
94.2%
Science
GPQA DiamondBest score 95.8%
90.9%
91.9%
91.7%
94.1%
Software engineering
Text Arena (Coding)Best score 1797.0
1598.61593.31581.61538.9
World knowledge
SimpleQA VerifiedBest score 75.6%
41.0%
34.2%
52.9%
66.2%

Model metadata