Google has begun allowing its employees to use Claude, the artificial intelligence model built by rival Anthropic. AI model developers including Anthropic now lead the industry to the point of raising calls to slow down development, yet Google's own model, Gemini, appears to have been pushed to the sidelines.
Google recently allowed its engineers to access Anthropic's coding AI models through an internal development tool, Business Insider reported on the 14th. Employees can use Anthropic's Opus 5 model through Antigravity, Google's in-house development platform.
Until now, Google had barred employees from using outside coding tools such as Anthropic's Claude Code or OpenAI's Codex, directing them to develop with Gemini instead. Only a small number of groups, including some teams at AI research subsidiary Google DeepMind and high-priority engineering projects, had secured permission to use Claude.
Business Insider described the move as an example of how intense competition in AI coding has become. Google, which develops Gemini, has invested billions of dollars in encouraging employees to use it, but shifted toward permitting use of a competitor's model as speeding up engineers' work became a priority.
"Engineers can use some third-party models in Antigravity," Google said. "Gemini remains the flagship and foundational model for internal development, and third-party models are provided on a quota basis to support specialized uses." The company added that employees can use Claude under per-user quotas and that the arrangement is designed to complement Gemini.
Some read the decision as a sign that Gemini's capabilities fell short enough to require allowing outside AI models internally. Google recently upgraded its Flash 3.8 model but still trails Anthropic and OpenAI in coding, and some Google employees have expressed frustration with Gemini's limitations, the outlet reported. Google has also released Gemini CLI, an open-source coding agent AI model, but it is seen as lacking the stable processing capability of Claude Code apart from its generous free capacity and large context window — the amount of text an AI model can hold in memory at one time.
Google is not the only Big Tech company to permit Claude. Amazon, which released its own AI model Nova, is also reported to have allowed employees to use Claude and Codex earlier this year. Both companies are investors in Anthropic. Google announced earlier this year that it plans to invest up to $40 billion in Anthropic. Amazon said earlier this year that it would invest up to an additional $25 billion in Anthropic on top of its existing $8 billion.
