Alibaba launches Qwen4-Max, priced at $1.5 per million input tokens
Alibaba's Qwen team released Qwen4-Max, the first model of the Qwen4 series, after a preview on November 10. It says the model is built on a new mixture-of-experts architecture with native vision and a 1M-token context window.
Claims (Alibaba):
- It outperforms Qwen3.8-Max with up to about 50% fewer tokens, and the average cost per task drops over 60% in Alibaba's internal evaluation. On Work360, token use falls about 24% (xhigh) and 35% (high), and the average cost per task goes from $4.43 to $1.67; on Terminal-Bench 4.0, token use falls about 48% and 49%.
- It is stronger with professional tools and at long-horizon execution, and Alibaba lists 3D modeling, game development, manufacturing, finance, law, healthcare and scientific discovery among the tasks it handles.
Pricing and access: $1.5 per million input tokens, $4.5 per million output tokens and $0.075 per million for implicit caching, all lower than Qwen3.8-Max, with caching down by 70%. The API is live on QwenCloud, and the model is available in Qwen Studio, QwenWork and Qoder and on the Token Plan.
Alibaba also says open weights of smaller Qwen4 models will be released.
Sources
Discussions
- Qwen 4 Max 居然和 Qwen 3.8 max 一样是 2.4 万亿的参数,这是。。。。训练了个寂寞吗?X · 25 comments
- Qwen4-MaxHacker News · 92 comments
- Qwen4-MaxHacker News · 163 comments
- Qwen4-MaxHacker News · 415 comments
Get news like this as a push.
AI Tracker alerts you when something changes for the tools you use.