Qwen3.7-Max

Qwen3.7-Max is a large language model developed by Alibaba's Qwen team, positioned as a coding and long-horizon task specialist that forgoes multimodal capabilities to maximize performance in its target domains . In independent benchmark testing by the ZangAI project, it scored 82.4 points and was characterized as having the most stable engineering capability among models tested, with "almost no major errors" and reliable graph rendering, though it lost points on visual and interaction consistency due to weak CSS variables and design systems . The model sits very close to the top tier: in ZangAI's evaluation, its gap to GLM 5.2 and Opus was under one percentage point, yet that marginal difference left it as "second place" in public attention while GLM captured the "domestic model #1" narrative during a period when Fable 5 was restricted . On cost efficiency, Qwen 3.7 Max was priced at roughly 23.5 RMB per test run in the same benchmark, substantially below Kimi K2.7-code at 164.6 RMB and Claude at 202.5 RMB . The model launched roughly a month before GLM 5.2, and its update frequency is described as slightly faster than GLM's .