Alibaba Launches 2.4T Qwen3.8-Max, Tops Kimi K3
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 flagship AI model with a 1-million-token context window, claiming it surpasses Kimi K3 on select benchmarks per Bloomberg.
- Alibaba plans to release open weights for both Qwen3.8-Max and the smaller Qwen3.8-27B model next week.
- Bloomberg framed Qwen3.8-Max's performance as rivaling global leaders including Anthropic's Fable, while The Information emphasized the new flagship is priced below Kimi K3.
- Reuters positioned the release within intensifying Chinese AI competition, noting DeepSeek is simultaneously driving costs lower.
- Qwen3.8-Max debuted at #4 on the Frontend Code Arena leaderboard with a score of 1,668, per its official account.
- Early preview tester Cole Brown said his hands-on experience was 'definitely not competitive with K3' and 'much less thorough,' tempering the headline performance claims.
- Unsloth announced plans to make Qwen3.8-27B runnable on GPUs under 20 GB, broadening local-deployment access.
Why it matters: Alibaba is shipping an open-weights model competing on top-tier benchmarks while explicitly undercutting Kimi K3 on price, per The Information — and Unsloth's 27B local deployment on sub-20GB GPUs means developers who couldn't afford enterprise hardware can now run a competitive model. The combination of open weights, benchmark parity with Anthropic Fable, and lower pricing is the exact combination that pressures proprietary frontier-model margins.
