|
1
|
Qwen 3.8 Max claims second place
Alibaba unveiled Qwen 3.8 Max, claiming it ranks second only to Anthropic's Fable 5 on global benchmarks — the highest any Chinese model has placed. The upgrade follows Moonshot's Kimi K3 release days earlier. The benchmark game has a new runner-up. Chinese labs are no longer chasing — they're competing.
|
|
2
|
Moonshot unveils largest open model
Moonshot AI released Kimi K3 — the world's largest open AI model, closing the gap on U.S. rivals. The model drops amid a wave of Chinese releases that analysts say mark a new phase in the China-U.S. AI race. Open-source scale just moved to China. The question is no longer whether Chinese models can compete — it's what open looks like when they do.
|
|
3
|
David Sacks warns on AI race
White House AI advisor David Sacks warned the U.S. could lose the AI race after Chinese models like Kimi K3 and Qwen 3.8 Max posted frontier-level results. The warning comes as Congress debates new China-tech export controls. Policy is waking up to what the benchmarks already showed. The regulatory response will shape who builds the next generation.
|
|
4
|
Apple Intelligence lands in China
Apple Intelligence received approval for launch in China through partnerships with Alibaba and Baidu, the two domestic AI players. The deal brings Apple's AI features to the world's largest iPhone market. Apple's AI play in China runs through Chinese models. The gatekeepers aren't in Cupertino.
|
|
5
|
Germany's Soofi S tops open-source
Germany's Soofi S may be the best-performing open-source AI model, according to early benchmarks. The model emerges as European AI efforts gain traction against U.S. and Chinese dominance in the open-weight space. Europe has a contender. The open-source map is no longer a two-country story.
|
