Topic
Alibaba
3 stories, newest first.
-
Alibaba releases Qwen4-Max, its new top AI model, and an outside test ranks it behind 13 U.S. models
Alibaba has released Qwen4-Max, the new top model in its Qwen family, which it says is built for coding, office work and long tasks. Developers can use it now, and so can people using Alibaba's apps. Alibaba says it cut the average cost of a task on its own work test by 62 percent from its last top model. Vals AI, an independent testing company, ranks it 14th of 42 models, behind 13 from U.S. companies, though its score is within the test's margin of the models just ahead.
-
Nous Research has released Arete-27B, a downloadable version of Alibaba's Qwen AI model that Nous says it trained further to stop repeating the Chinese government's account of sensitive topics. On Nous's own 803-question test, graded by a different AI model, the share of answers that took the state's line fell from 22% to 0.9%. Nous ran the original and new models with different settings from the other models it tested, and says it has not fixed every behavior tied to the original model's views.
-
Apple says the first of its Apple Intelligence features, from live translation to email summaries, reach iPhones in mainland China in December, after app developers can try them from Nov. 4. Alibaba's Qwen AI will handle some Siri and writing requests there. Apple has not said when Siri AI, its new version of Siri, or the features for its other devices will follow.