Alibaba Group's Qwen3.8-Max model operated autonomously for 16 days, completing 265 commits on a public Git repository without human intervention. This extended operational period challenges the typical failure horizon of unattended coding agents, which is usually measured in hours. The success is attributed not solely to the model's intelligence but to the surrounding scheduler architecture that maintained its operation. AI
IMPACT Demonstrates advanced agentic capabilities and scheduler architecture, potentially extending autonomous operation times for AI models.
RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
