Tencent Hunyuan
#24—
CHINAOpen SourceConfidence: LOW
Tencent's AI division building open-source Hunyuan image and video models
Website →Founded 2023
0
Overall
0
Overall
Text
60
Coding
45
Reasoning
40
Image
72
Video
75
Multimodal
55
Agents
28
About
Tencent's AI division building open-source Hunyuan image and video models
Pillar Breakdown
Why They Moved
- Hunyuan Image 3.0
- HunyuanVideo Release
Pillar Scores
Capability62
Velocity55
Adoption & Distribution55
Compute & Efficiency55
Ecosystem & Partnerships52
Trust / Safety / Access52
Timeline
🚀
Sep 20, 2025
Hunyuan Image 3.0
Open-sourced the largest image model at 80B parameters with MoE+Transfusion architecture.
🚀
Dec 3, 2024
HunyuanVideo Release
Tencent open-sources HunyuanVideo, a 13B parameter video model becoming foundation for many fine-tunes.