Alibaba's latest AI model ran autonomously for 35 hours to optimize code for its own custom chip
35 hours. That's how long Alibaba's Qwen3.7-Max ran autonomously optimizing code for its own chip—and it's now matching Claude Opus on benchmarks.
Why it matters
Alibaba's Qwen3.7-Max demonstrates a new frontier in autonomous agent capability and long-context reasoning, positioning Chinese AI labs as competitive with frontier Western models on reasoning tasks while achieving practical chip-design applications.
The key facts
7 to knowQwen3.7-Max matches Claude Opus 4.6 on benchmarks
Model ran autonomously for 35 hours optimizing code for custom chip
Beats DeepSeek V4 Pro and Kimi K2.6 on tested benchmarks
Proprietary model built for long-running autonomous agent tasks
Demo includes four-legged robot control capability
Released by Alibaba's Qwen team
Published May 23, 2026
Go to the source
The Decoderthe-decoder.com
Publisher excerpt: Alibaba's Qwen team releases Qwen3.7-Max, a proprietary model built for long-running autonomous agent tasks. It matches Claude Opus 4.6 on benchmarks and beats Chinese rivals like DeepSeek V4 Pro and Kimi K2.6. The team also demos the model steering a four-legged robot.