#14
GLM 5.2 Fast
glm-5.2-fast
0.0
Helpfulness
Instruction Following
Comprehension
Empathy
Creative Writing
Helpfulness
0.0
Instruction Following
0.0
Comprehension
0.0
Empathy
0.0
Creative Writing
0.0
Speed
Avg 88 tok/s
Release Date
June 23, 2026
Lab
Zhipu AI
Type
Open Source
Context Size
1M
Max Output Tokens
131.1K
Cost per 1 million tokens
$2.10 / $6.60
Model Inputs*
Text
Model Outputs*
Text
Tool Calling*
Enabled
Overall Assistant Score
An average score combining the 5 main categories.
90.10 pts
Rank #14
81th Percentile
0
Novice
33
Capable
66
Proficient
100
Expert
GLM 5.2 Fast is the snappy serving of GLM 5.2—same open assistant brain, tuned for lower latency when you are chatting interactively. Our tests find nearly the same helpfulness and instruction following as standard 5.2, with a small trade in depth on the hardest multi-step asks. Empathy and creative writing stay in the same utilitarian band. Choose Fast when response time matters in a live conversation; stay on standard 5.2 when you can wait for a more deliberate finish.

Overall Score vs Price

Price axis is logarithmic (USD per 1M input tokens) · Higher score and lower price is better

Ringed point is this model

Anthropic
DeepSeek
Google
Meta
Moonshot
OpenAI
SpaceX AI
Zhipu AI

Intelligence

Overall Score · Higher is better

Speed

Output tokens per second · Higher is better

Price

USD per 1M input tokens · Lower is better

Zhipu AI Models

Overall Score · Same provider

GLM Family

Overall Score · Same model family

Closest Rivals

Overall Score · Nearest by overall rank