Qwen4 Max API is out; researcher notes big CoT reduction, eyes a 27B variant

Xianbao_QIAN · x · 2026-11-16

Xianbao QIAN notes the Qwen4 Max API is live and shows a big "intelligence density" win — noticeably shorter chains of thought. Unsure where the CoT reduction comes from, he's looking forward to a coming 27B variant trained on the same data/method, which could be turbo-charged via domain adaptation and local deployment.

Original post →

More from Models

Models channel →