Qwen3.8-Flash-Next
Alibaba Qwen · · 18 days after Qwen-Image-3.0
Partially verified
0 of 1 published values traced to a source2 primary sources, 2 archived2 not disclosed by the lab
Why this status
Cited to the lab's own announcement and the official model card. Not verified because the release date could not be found in any cited primary source — the value may be correct, but this dataset cannot yet show where it came from. Modalities from the model card's own pipeline_tag "image-text-to-text" on Qwen/Qwen3.8-Flash-Next.
What these states mean · How records are judged · Has this record changed?
Qwen introduced Qwen3.8-Flash-Next as the first open-weight release under a new architecture aimed at cost efficiency, combining Gated DeltaNet and Gated Attention with Qwen Sparse Attention and a Gated Residual stream. Qwen states a 125B-parameter main model supplemented by an additional 51B N-gram embeddings, with 6B parameters activated per token, and reports training cost around one ninth of Qwen3.7-Plus.
Details
| Released | August 3, 2026 |
|---|---|
| Company | Alibaba Qwen |
| Family | Qwen |
| Released alongside | Qwen3.8-Max |
| Era | Agentic systems |
| Type | Language model · LLM |
| Modalities | Text + Image → Text |
| Capabilities | Not recorded |
| Weights | Open weights |
| Licence | qwen-community-1.0 |
| Context window | Not disclosed |
| Parameters | Not disclosed |
Current information, elsewhere
This record is a historical one: every figure above is traced to a dated primary source. For what is true today — live pricing, current rankings, availability — these are better maintained than anything here could be.
- Current price and providers OpenRouter ↗
- Current benchmark scores Artificial Analysis ↗
- Weights and model card Hugging Face ↗
Sources
Record status: partly verified · confidence 60/100 · 2 of 2 primary
Cited to the lab's own announcement and the official model card. Not verified because the release date could not be found in any cited primary source — the value may be correct, but this dataset cannot yet show where it came from. Modalities from the model card's own pipeline_tag "image-text-to-text" on Qwen/Qwen3.8-Flash-Next.
- qwen.ai Official announcement archived
- huggingface.co Model card archived
Primary means published by the organisation that made the model. Verified records require at least one.
Qwen family
- Qwen1.5-72B February 4, 2024
- Qwen2-72B June 7, 2024
- Qwen2.5-72B September 19, 2024
- QwQ-32B March 6, 2025
- Qwen3-235B-A22B April 29, 2025
- Qwen3-30B-A3B April 29, 2025
- Qwen3-Coder-480B-A35B-Instruct July 22, 2025
- Qwen3-Max-Thinking January 23, 2026
- Qwen3-Coder-Next February 2, 2026
- Qwen3.5-397B-A17B February 14, 2026
- Qwen3.6-Plus April 2, 2026
- Qwen3.6-35B-A3B April 15, 2026
- Qwen3.6-27B April 21, 2026
- Qwen3.7-Max May 16, 2026
- Qwen3.7-Plus May 21, 2026
- Qwen3.8-Max August 3, 2026
- Qwen3.8-Flash-Next August 3, 2026
- Qwen3.8-27B August 5, 2026
- Qwen3.8-2.4T-A95B August 8, 2026
- Qwen3.8-Omni-Flash September 18, 2026
Compare Qwen3.8-Flash-Next with another model →
See Qwen3.8-Flash-Next on the interactive timeline →