Qwen3.8-Max-0902 tops CodeArena coding leaderboard at ~$5 per million tokens
千问大模型 · wechat · 2026-09-02
Alibaba's Qwen team shipped Qwen3.8-Max-0902, post-trained for coding and professional office work based on a month of enterprise feedback:
- Climbed 22 points to 1691 on the CodeArena frontend leaderboard, taking first place, with claimed gains in multi-step reasoning, tool calling, and end-to-end app generation
- Strong price-performance: $5 per million tokens average, vs $20 and $12 for the #2 and #3 models on the leaderboard
- Live on the Qwen AI platform API and integrated into Qwen Office, Qoder, and the Qwen app
Related event: Alibaba Releases Qwen3.8-Max-0902 With 2.4T Parameters and 1M Context(4 posts)→
More from Models
- 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
- A quick test for benchmaxxed demos: swap in Mirror's Edge and see if it generalizes — teortaxesTex · 2026-10-18
- Did OpenAI distill DeepSeek's gains? Benchmark shows OpenAI's Luna scoring just 38 — teortaxesTex · 2026-10-18