Alibaba Tongyi Qwen3.8-Flash Open-Sourced: 125B-Parameter MoE Model with Training Cost Only 1/9 of Predecessor
Summary
Alibaba's Tongyi Qianwen team released Qwen3.8-Flash, a multimodal MoE model with 125B total parameters and 6B activated per token, supporting 262K context expandable to 1M. The training cost is only one-ninth of the predecessor Qwen3.7-Plus, yet it performs better on coding and office tasks. Model weights are open-sourced, with API pricing at 1 yuan per million input tokens and 3 yuan per million output tokens.
Bubble analysis
This news is mainly about a model release and technical details, with weak direct relevance to the AI investment bubble. Although lower training costs could affect the cost structure of the AI industry, it does not provide specific investment amounts or capex data, so its reference value for bubble assessment is limited.
Related signals
linked by shared tags
- Bilibili Open-Sources the Index-Translate Multilingual Translation Model Family, Text Models Cover 150 Languages#open-source#qwen
- Amazon releases its own Jev clone as decision models flood the web#model-release
- Google releases Gemini 4 Argon, called its most powerful model yet#model-release
- Google announces Gemini 4 Argon AI model, but you can't use it yet#model-release
- After Alibaba's 20GW Call: Five Shifts in the Compute Race#alibaba