Google has expanded access to Anthropic’s Claude to engineers across the company, allowing them to use one of its strongest coding models for internal development. The move marks a notable shift for Google, which has traditionally pushed employees to use its own Gemini models and restricted access to external AI coding tools.
Claude Access Expands Across Google
Google recently made Anthropic’s Opus 5 model available to engineers across the company through its internal development platform, according to people familiar with the matter and a Google spokesperson.
However, employees cannot freely use Claude through any external service. Access is provided through AntigravityGoogle’s internal version of its development platform, and is subject to individual usage quotas.
Google said third-party models are intended to complement, rather than replace, Gemini.
Gemini Remains Google’s Primary AI
A Google spokesperson said Gemini remains the company’s primary and foundational model for internal development.
The company is nevertheless allowing selected third-party models for specialised use cases. Claude’s inclusion is significant because Google has historically restricted most employees from using external coding tools such as Anthropic’s Claude Code and OpenAI’s Codex.
Previously, Claude access had largely been limited to Google DeepMind teams and certain employees working on high-priority engineering projects.
The broader rollout suggests Google is becoming more pragmatic about which AI tools can make its engineers more productive.
Coding Competition Is Driving the Change
The decision comes amid an increasingly intense competition between Google, Anthropic and OpenAI over AI-powered software development.
Google has invested heavily in Gemini and has encouraged employees to use its internal AI systems. However, some engineers have reportedly found Anthropic and OpenAI’s coding tools more capable for certain complex programming tasks.
That frustration had created a divide within Google, particularly after DeepMind engineers were allowed to use Claude while many other Google engineers could not.
By opening access more broadly, Google can allow engineers to select the model that works best for particular development tasks while retaining Gemini as its default system.
Google Is Also an Anthropic Investor
The decision is particularly notable because Google is both a competitor to and investor in Anthropic.
Google has committed to invest up to $40 billion in Anthropicgiving it a financial interest in the success of a company whose AI models increasingly compete with Google’s own technology.
The relationship illustrates how the AI industry is becoming more interconnected. Technology companies can simultaneously compete with AI labs, invest in them and use their models internally.
A New Approach to AI Development
Google’s move also reflects a broader change in how major technology companies approach AI tools. Instead of relying exclusively on internally developed models, companies are increasingly giving employees access to multiple systems and allowing them to choose specialised models for specific tasks.
Amazon has similarly expanded employee access to Claude and OpenAI’s coding tools.
For Google, the decision could help engineers work more efficiently while giving the company additional insight into how competing AI systems perform in real-world software development.
Summary: Google has expanded access to Anthropic’s Claude Opus 5 for engineers across the company through its internal Antigravity development platform. Gemini remains Google’s primary AI model, but employees can now use Claude for specialised coding tasks under usage quotas. The move highlights the intensifying AI coding competition and Google’s willingness to use a rival’s technology despite developing its own Gemini models.