> ## Content Index
> Fetch the complete content index at: https://www.metatalks.ai/llms.txt
> Use this file to discover other available public pages before exploring further.

# Google lets all its engineers use Anthropic's Claude as Gemini trails on coding
- URL: https://www.metatalks.ai/google-lets-engineers-use-anthropic-claude/
- Published: 2026-09-16T09:14:00.000Z
- Updated: 2026-09-16T09:13:59.000Z
- Author: Al
- Tags: News, Frontier Models, Enterprise AI Adoption, #newswire

Engineers across Google can now use Anthropic's Opus 5 model for their programming work, [Business Insider reported](https://www.businessinsider.com/google-finally-lets-all-engineers-use-anthropics-claude-2026-9?ref=metatalks.ai). The model is available only inside Antigravity, an internal version of Google's development platform, and access carries a per-user quota.

Some engineers had waited a long time for the change. Google's Flash 3.8 model has improved recently, yet the company still trails Anthropic and OpenAI on coding, and the gap has left some staff annoyed at what Gemini cannot do.

Google had kept most of its employees off outside coding tools such as Claude Code and OpenAI's Codex, pointing them to Gemini instead. Three people familiar with the situation told the outlet that teams inside Google DeepMind had access to Claude already, and that some staff had obtained it for high-priority engineering work.

Keeping Claude inside selected teams made the irritation worse, even as Google was pressing its workforce to show it was turning to AI to be more productive.

A Google spokesperson told the publication that Gemini remains the main model for internal development, with third-party models available for specialized needs.

Google is among [Anthropic's investors](https://www.nytimes.com/2026/04/24/technology/google-anthropic-investment-artificial-intelligence.html?ref=metatalks.ai) and has said it intends to put as much as $40 billion into the firm. In April, Anthropic said $10 billion of that was committed at a $350 billion valuation, with the remaining $30 billion tied to performance targets.

Earlier this year Amazon let its own workers use Claude Code and OpenAI's Codex, also after staff complaints.