Alibaba Releases 2.4T-Param Qwen3.8-Max
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Alibaba released Qwen3.8-Max, a 2.4 trillion-parameter model with a 1 million-token context window designed for advanced coding, real-world tasks, in-depth research, and long-horizon challenges.
- Alibaba claims Qwen3.8-Max outperforms Moonshot's Kimi K3 on some benchmarks and positions it alongside global leaders including Anthropic's Fable.
- Alibaba plans to release open weights for both Qwen3.8-Max and Qwen3.8-27B next week, with SCMP noting the flagship was made widely accessible ahead of that release.
- The Information reports Alibaba is pricing the new flagship below Kimi K3, framing the launch as a cost offensive in China's tightening model race.
- Alibaba framed the release as benchmark-competitive with GPT-5.6 Sol and Claude Fable 5 per Neowin, while the model hit #4 on the Frontend Code Arena leaderboard with a score of 1,668.
- Early reactions were split: some testers praised the 27B variant's ability to run on GPUs under 20GB, while others (e.g. @dtcb) said the preview felt 'not competitive with K3' in thoroughness despite its linear-attention speed.
Why it matters: By undercutting domestic rival Kimi K3 on price and shipping open weights next week, Alibaba is competing on accessibility and cost against both Chinese labs and frontier Western players like Anthropic. One community observer noted the 27B variant running on sub-20GB GPUs puts it 'well within reach of a LOT of everyday users,' highlighting the open-weights release as the real distribution lever behind the flagship headline.
Ask SkimNews
