Alibaba's Qwen3.7-Max AI model ranks 4th in global coding benchmark

Alibaba's new AI model, Qwen3.7-Max, has achieved the fourth position on the Code Arena leaderboard, surpassing models from OpenAI and Google. This specialized coding agent is the only non-US developed model to rank in the top five, which are otherwise dominated by Anthropic's Claude models. Code Arena evaluates models on their ability to build complete web applications from scratch based on user prompts, with real-world developer preferences influencing the rankings. AI

IMPACT Highlights the growing capability of specialized AI coding agents and signals increasing competition from non-US developers in this domain.

RANK_REASON A non-US AI model achieved a top-5 ranking on a specialized coding benchmark, surpassing major US competitors. [lever_c_demoted from significant: ic=1 ai=1.0]

Read on SCMP — Tech →

• Alibaba
• Anthropic
• Carnegie Mellon University
• Claude
• Code Arena
• Google
• OpenAI
• Qwen3.7-Max
• University of California, Berkeley
• University of California San Diego

• model release
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →