Google lets engineers use Anthropic's Claude for coding work

Google has granted its internal engineers permission to use rival artificial intelligence (AI) model Claude from Anthropic to assist with coding tasks, easing previous restrictions that required employees to rely almost exclusively on its in-house Gemini AI.

Engineers across the tech giant can now access Anthropic's Claude Opus 5 model for internal work via Antigravity, Google's platform designed for software development and AI agent workflows.

Access to Claude was previously restricted to the Google DeepMind team and select engineering groups working on high-priority projects.

The shift follows long-standing requests from staff who cited operational limitations when relying solely on Gemini, with workers expressing hope that exposure to competing systems will help Google catch up to rivals Anthropic and OpenAI in AI-assisted coding capabilities.

Google previously enforced strict controls on third-party coding tools, including Claude and OpenAI Codex, to encourage internal adoption of Gemini.

The move highlights the deepening relationship between the two firms, following Google's announcement last year of plans to invest up to US$40 billion in Anthropic.

"Gemini remains our primary model for corporate development," a Google spokesperson said, adding that staff have been provided usage quotas for Claude to complement existing Gemini workflows.

Google is not alone in adapting its internal policies, as employees at e-commerce titan Amazon have also begun lobbying management to permit third-party AI models for software development.

The policy change underscores a broader shift in corporate software development, where technology majors increasingly adopt multi-model strategies rather than relying on a single proprietary AI system.

Source: Bussiness Insider Africa