Anthropic said it is operating an internal AI model, Model 2, that performs better than its top publicly known model, Claude Mythos 5, but it does not plan to release it. In an AI risk report, the company disclosed safety review results for Mythos 5 and Model 2 and described Model 2 as slightly more capable. Employees also rated it as improved on many tasks, though gains were limited. Benchmark and safety test results were mixed.

Alibaba has teased its Qwen3.8 model, calling it the most powerful in its Qwen line and claiming it ranks second only to Anthropic’s Claude Fable5. It previewed Qwen3.8-Max-Preview at the World Artificial Intelligence Conference in Shanghai, describing a multimodal model with 2.4 trillion parameters. Alibaba released no benchmark scores, model card or activation-parameter details, and the model has not been independently verified. The preview is available via subscription and developer platforms at discounted trial pricing.