Alibaba
China · China · 4 model families / 1 agent tools
Headquartered in China. This index covers 4 model families / 17 versions and 1 agent tools in Language & multimodal / Speech & audio / Image / Video, 6 of them active and 8 with open weights. Lineups are listed by category below, with a full lineage ordered by official release date.
Families4Versions17Agents1CategoriesLanguage & multimodal / Speech & audio / Image / VideoActive / open6 / 8Bench records1Updated2026-11-08
01 / Models
Language & multimodal
1 entries| Model | Status | Openness | versions | Last updated | |
|---|---|---|---|---|---|
| Qwen | Active | Open weights | 3 |
02 / Models
Image
1 entries| Model | Status | Openness | versions | Last updated | |
|---|---|---|---|---|---|
| Tongyi Wanx (Image) | Active | Closed source | 5 |
03 / Models
Video
1 entries| Model | Status | Openness | versions | Last updated | |
|---|---|---|---|---|---|
| Wan Video | Active | Closed source | 6 |
04 / Models
Speech & audio
1 entries| Model | Status | Openness | versions | Last updated | |
|---|---|---|---|---|---|
| Tongyi Speech (CosyVoice) | Active | Open source | 3 |
05
Agents
06
Version lineage
- Wan 3.0 · Wan Video Video TBD
- Wan 2.7 · Wan Video Video Superseded TBD
- wan2.6-t2i · Tongyi Wanx (Image) Image TBD
- Fun-CosyVoice 3.0（0.5B） · Tongyi Speech (CosyVoice) Speech & audio Third generation, open-sourced December 2025; improves content consistency, speaker similarity and prosody over 2.0, with 9 languages and 18+ Chinese dialects per the HF model page and project materials.
- Wan 2.6 · Wan Video Video Superseded TBD
- Wan 2.5 · Wan Video Video Superseded TBD
- wan2.5-t2i-preview · Tongyi Wanx (Image) Image Superseded TBD
- Wan 2.2 · Wan Video Video Superseded TBD
- Qwen3 · Qwen Language & multimodal TBD
- Wan 2.1 · Wan Video Video Superseded TBD
- wanx2.1-t2i (plus/turbo) · Tongyi Wanx (Image) Image Superseded TBD
- CosyVoice 2.0（CosyVoice2-0.5B） · Tongyi Speech (CosyVoice) Speech & audio Superseded Released December 2024 with a streaming-first redesign and ~150 ms first-packet latency per project materials.
- Qwen2.5 · Qwen Language & multimodal TBD
- CosyVoice 1.0 · Tongyi Speech (CosyVoice) Speech & audio Superseded First released July 2024 alongside SenseVoice under the FunAudioLLM project.
- wanx-v1 · Tongyi Wanx (Image) Image Superseded TBD
- 通义万相（首发） · Tongyi Wanx (Image) Image Superseded TBD
- Qwen3-235B-A22B-Instruct-2507 · Qwen Language & multimodal TBD