Zhipu AI
#14—
CHINAOpen SourceConfidence: MEDIUM
Tsinghua-spinoff building the GLM model family, CogView image, and CogVideo models — GLM-5 approaches Claude Opus on coding
Website →Founded 2019
0
Overall
0
Overall
Text
78
Coding
70
Reasoning
62
Image
70
Video
68
Multimodal
65
Agents
50
About
Tsinghua-spinoff building the GLM model family, CogView image, and CogVideo models — GLM-5 approaches Claude Opus on coding
Pillar Breakdown
Why They Moved
- GLM-5 Launch
- GLM-4.7 Open-Source Release
- CogView4 Open-Source
- CogVideoX-5B Release
- Zhipu Raises $400M
Pillar Scores
Capability78
Velocity82
Adoption & Distribution60
Compute & Efficiency62
Ecosystem & Partnerships58
Trust / Safety / Access65
Timeline
🚀
Feb 11, 2026
GLM-5 Launch
745B MoE flagship with 44B active params, approaching Claude Opus on coding benchmarks. Trained entirely on Huawei Ascend chips, MIT licensed.
🚀
Dec 23, 2025
GLM-4.7 Open-Source Release
Open-source coding and agent model rivaling GPT-4 on code, MIT licensed.
🚀
Mar 18, 2025
CogView4 Open-Source
First open-source model generating Chinese characters in images, Apache 2.0 licensed.
🚀
Sep 15, 2024
CogVideoX-5B Release
Open-source 5B video model outperforming VideoCrafter-2.0, published at ICLR 2025.
💰
May 28, 2024
Zhipu Raises $400M
Zhipu raises $400M at ~$3B valuation from Prosperity7 Ventures.