Alibaba Qwen3.8-Max-0902 Tops Code Arena WebDev and Cost-Performance Frontier

Qwen ·

Key Info

Alibaba's Qwen3.8-Max-0902 model has claimed the #1 spot on Code Arena's WebDev leaderboard with 1,691 points, while becoming the highest-scoring model on the cost-performance Pareto frontier at a blended $5/MToken.

Highlights

  • On the updated frontier, Qwen3.8-Max-0902 outperforms HY4 Preview (1,629 pts at $2.08/MToken) and Qwen3.8-Flash-Next (1,620 pts at $0.39/MToken), balancing a higher score with reasonable cost.
  • Other top leaderboard models such as Claude Opus 5 (Max), ranked #2 overall, have dropped off the Pareto frontier despite keeping top positions, leaving Qwen3.8-Max-0902 as the frontier's standout.
  • The upgraded model, built with 2.4T parameters and a 1M-token context window, is post-trained for coding and collaborative agentic workflows, reinforcing its strength in agentic coding tasks.
Loading...