Qwen3.7 models

Strongest: Qwen3.7-Max

Alibaba superseded Qwen3.7-Max with Qwen3.8-Max on 2026-08-03 and now lists qwen3.7-max among legacy models it no longer recommends; it stays GA and billable at the same price, while its successor costs less ($2.00/$6.00) and ships open weights. Qwen3.7-Max is Alibaba Cloud's proprietary flagship — a deliberate break from the Qwen line's open-weights tradition — released 2026-05-21 with a 1M-token context window (up from 256K on Qwen3.6-Max), thinking mode on by default, and a June 10 snapshot that added visual understanding for "multimodal interactive hybrid agent" work: reading screens, operating GUIs, and coding from visual references. At $2.50/$7.50 per 1M tokens it undercuts the Western flagships above it, and it ranked among the top proprietary models on the Artificial Analysis Index (46, #13 of 162 tracked as of 2026-07-03, before the index was rescaled). Buyers who chose Qwen for its Apache-2.0 freedom should note: this one is closed, API-only, and sold through Alibaba Cloud Model Studio.