Alibaba Unveils Qwen3.8-Max AI Model to Challenge Anthropic's Claude
Alibaba/Qwen
Anthropic
Alibaba has introduced a new AI model, Qwen3.8-Max, capable of autonomous operation for up to five days. The company claims the model can independently handle tasks in programming, scientific research, and data analysis, and it has shown strong performance in an online competition. These capabilities rival those of Anthropic's Claude models, though independent verification is still pending.
Chinese tech giant Alibaba has presented its new AI model, Qwen3.8-Max, which the company says can work without human intervention for several days, potentially up to five days. During an experiment, the model ran continuously for 125 hours, autonomously reproducing an experiment from a mathematical scientific paper and then suggesting improvements. Alibaba also tested the model in an online competition against 526 human teams, where Qwen3.8-Max outperformed most participants within 24 hours, losing only to 68 teams. According to Bloomberg, the announced capabilities are comparable to those Anthropic is developing in its Claude models, such as independently analyzing code, finding errors, planning fixes, and completing long-duration tasks with minimal user intervention. Anthropic previously stated its Claude Sonnet 4.5 model can work autonomously for over 30 hours. Alibaba claims Qwen3.8-Max can operate in a similar mode, but neither company's test results have been independently verified yet.
- Abbreviations
- LLM = Large Language Model — большая языковая модель
Source: Rusbase (RB.RU) —
original