The new model, Claude Opus 4.5, was more capable than previous versions at software engineering work, like fixing bugs, without needing a user’s help, Anthropic said. Opus 4.5, released on Monday, is also intended to be better at carrying out complicated multistep tasks on a user’s computer and the internet.
Founded in 2021 by former employees of OpenAI, San Francisco-based Anthropic now has more than 300,000 business customers who use its models to streamline workplace tasks – particularly in the field of computer programming, where the start-up has emerged as a market leader.
Anthropic’s new model has hit a new coding milestone of sorts, according to Scott White, head of product for Claude AI models. Opus 4.5 was the first to score higher than any of the company’s human candidates on a challenging take-home engineering assignment that prospective employees were asked to complete, he said.
