Alibaba ships Qwen 3.6 Plus and Qwen 3.6 Max Preview — agentic push in two-week tempo
Qwen 3.6 Plus dropped April 2; Qwen 3.6 Max Preview followed April 20. Alibaba's framing: "accelerating agentic AI deployment for enterprises and Alibaba's AI applications." Built on the Qwen 3.5 native-multimodal foundation from February, which supports 201 languages.
The cadence is the story. Qwen 3.5 and Qwen 3.5 Plus shipped February 16; Qwen 3.5-122B-A10B, 35B-A3B, and 27B on February 24; the smaller 0.8B/2B/4B/9B variants through March; Qwen 3.6 Plus on April 2; and Qwen 3.6 Max Preview on April 20. That's a full generation moved in roughly nine weeks — faster than most Western frontier labs publish on, and with significantly more model-size diversity per release.
Architecturally, Qwen 3.5 introduced native multimodal training (text + image + video in one system) and pushed language coverage to 201 languages — up from the prior generation's 82. The Qwen 3.6 cycle's pitch is agentic: tool-use, workflow chaining, and integration with Alibaba's enterprise stack.
This matters for the China AI frontier coverage we'll keep adding. Qwen, DeepSeek, Moonshot/Kimi, ByteDance/Doubao, and the Baidu/Tencent stack are not lagging the US frontier in 2026 — they're shipping different shapes of capability on a different cadence.