Z.ai (Zhipu)Language modelSupersededOpen weights
GLM-4.5
MIT-licensed agentic model that made Chinese open weights a coding default.
Overall Intelligence
#47
of 56
24.8 pts behind #1
0
races led
0
weeks at #1
63.9
score
- Released
- Jul 28, 2025
- 1.2 yr ago
- Context
- 128K
- 96K max output
- Price
- $0.6 / $2.2
- per 1M in/out · $1.00 blended
- Speed
- 70 tok/s
- 1.2s to first token
- Modalities
- text
- outputs text
- Weights
- Open
- 355B MoE (32B active)
The 30-second version
MIT-licensed agentic model that made Chinese open weights a coding default.
Key takeaways
- 1MIT-licensed agentic model that made Chinese open weights a coding default.
Where it ranks
Current snapshot · illustrative scores · click any race to watch it
GLM-4.5Best on each axis
Overall
Coding
47Coding+24.9 pts48Frontend & UI+155 Elo47Backend & APIs+22.4 pts47Distributed Systems+25.9 pts45Full-Stack Apps+22.3 pts47Debugging & Repair+25.4 pts46Agentic Coding+25.4 pts48Code Review & Security+23.5 pts50Data Science & SQL+25.5 pts43Mobile Apps+19.7 pts45Terminal & DevOps+27.3 pts46Competitive Programming+716
Hard Reasoning
Math
Science
Writing & Language
Multimodal
Tool Use & Agents
Speed & Cost
Overall Intelligence over time
#1 frontierGLM-4.5V4.1 FlashGLM-5.3
Update history
The distilled version first; open any release for the deeper breakdown.
The GLM family