Alibaba says its newest large language model, Qwen3.8-Max, is now the second-strongest model in the world, trailing only Anthropic's Claude. The claim was reported by Startup Fortune, surfaced via Google News.
That is the extent of what the source establishes. The report, as summarized, does not specify which benchmarks or evaluations Alibaba used to arrive at the ranking, how wide the gap to Claude is said to be, or which Claude model the comparison targets. Nor does it detail pricing, availability, or how the model was trained.
Those gaps matter, because the claim is Alibaba's own. Model makers routinely publish self-selected benchmark results alongside a launch, and independent testers and third-party leaderboards often produce a different ordering once they run their own evaluations. Until that outside verification arrives, "second only to Claude" is a vendor's framing rather than a settled fact.
Still, the framing itself is the news. Alibaba is not positioning Qwen against other Chinese labs or against a broad field — it is naming a single American company as the only one ahead of it. That is a statement about where the company believes the frontier of AI now sits, and about who it considers the benchmark to beat.
It matters because the pecking order at the top of the AI market shapes which models companies build on, and a Chinese lab publicly measuring itself against Anthropic signals how tight that race has become.