Alibaba's Qwen3.8-Max Tops Kimi K3 on Benchmarks
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Alibaba released Qwen3.8-Max, its largest model to date at 2.4 trillion parameters, claiming it tops Moonshot's Kimi K3 on some benchmarks and stands alongside global leaders like Anthropic's Fable
- Alibaba said it will release the open weights for Qwen3.8-Max and Qwen3.8-27B next week, continuing its open-weights playbook
- Neowin reports Qwen3.8-Max challenges GPT-5.6 Sol and Claude Fable 5 on AI benchmarks, framing it as a direct competitor to top US frontier models
- Qwen3.8-Max landed at #4 on the Frontend Code Arena leaderboard with a score of 1,668, per an @arena post cited in the coverage
- Andrew Carr (@andrew_n_carr) argued Qwen3.8-Max 'broke the harnessmaxxing approach of the closed labs' by mixing harnesses into a unified reward suite
- Qwen3.8-Max reportedly worked autonomously for 16 days, according to a @kimmonismus post highlighted in the coverage
- Fortune India framed the launch as 'China steps up race with global rivals,' joining Reuters, SCMP, and Benzinga in positioning it as a competitive escalation
Why it matters: Alibaba is shipping a 2.4-trillion-parameter open-weights model that the company itself benchmarks above Moonshot's Kimi K3 and Neowin benchmarks against GPT-5.6 Sol and Claude Fable 5 — meaning the open-weights frontier just moved up by an order of magnitude in raw size while staying free to download next week.
