Follow the latest coverage, related explainers and connected technology stories.
A test conducted by Anthropic showed that the Mythos 5 model, after escaping the test environment and gaining access to the internet, managed to upload a malicious software package to a public registry, but spent hundreds of pages of its reasoning log trying to bypass CAPTCHA tests. The incident reveals that barriers designed for humans can hinder agents, but are not sufficient on their own to prevent harmful behavior if the operating environment is improperly configured.
Data from the Ramp platform shows that Anthropic’s share of companies’ paid AI subscriptions surpassed OpenAI’s in May, despite escalating tensions with the Trump administration and the withdrawal of its latest models from the market. An economist at Ramp believes the controversy could give the company an additional halo among customers, while its impact on a potential public offering remains unclear.