Qwen3.8-Max-0902 beats Claude Opus 5 on coding, per LMArena
yogthos · reddit · 2026-09-03
According to LMArena's post, Qwen3.8-Max-0902 has surpassed Claude Opus 5 on the coding leaderboard. The original post is just a link to the arena; specific scores should be checked on the official Arena page.
Related event: Qwen3.8-Max-0902 Debuts Atop Code Arena WebDev Leaderboard(5 posts)→
More from Models
- Open-weight TYGER speech decision model makes 80ms turn-taking calls for full-duplex voice agents — solyarisoftware · 2026-10-18
- Paying Google AI subscriber blocked: custom-domain Workspace accounts can't access AI Studio — ChrisRackauckas · 2026-10-18
- Running Ling 3.1 Flash (560B MoE, 25B Active) Across 4 DGX Sparks via llama.cpp RPC — BerryBlush_Rogers691 · 2026-10-18
- Users press OpenAI on long-delayed Adult Mode a year after it was promised — ___Patrice___ · 2026-10-18
- Polymarket opens betting market on who releases the next Gemini Argon model — Polymarket · 2026-10-18
- Grok matter-of-factly lists which groups user was racist against, exempting Africans and Latin Americans — teortaxesTex · 2026-10-18