Alibaba released Qwen4-Max, with a parameter scale of 2.4 trillion
On November 16, Alibaba released Qwen4-Max. This model has a total parameter count of 2.4 trillion, supports native visual capabilities, and has a context window of 1 million tokens; each token activates 71 billion parameters. Alibaba claims that Qwen4-Max ranks sixth in Code Arena and has achieved results in benchmark tests such as Agents' Last Exam, Vals Index, JobBench, and Chartography.
Alibaba disclosed that in its Work360 benchmark test, Qwen4-Max outperformed the previous generation Qwen3.8-Max overall; in high-intensity inference mode, token usage decreased by 35%, and the average cost per individual task was reduced by about 60%. The model is available through QwenCloud's Token Plan and API, and can be used in QwenWork and QwenStudio; the input price is $0.075 per 1 million tokens. Alibaba stated that smaller-scale open-weight models from the Qwen4 series will be released to developers in the future.
ChainCatcher reminds readers to view blockchain rationally, enhance risk awareness, and be cautious of various virtual token issuances and speculations. All content on this site is solely market information or related party opinions, and does not constitute any form of investment advice. If you find sensitive information in the content, please click , and we will handle it promptly.






