More from Sourceable
Continue reading our latest insights
Continue reading our latest insights
Google has spent years building and promoting Gemini.
Now, its engineers can also use a major rival.
As of September 15, 2026, Google has expanded Claude access through its Antigravity development platform. Claude comes with usage limits. Gemini remains a major part of Google's AI strategy.
The move is important for the AI coding market.
It shows that Google wants engineers to have more model choice. It also shows how important AI coding agents have become for large software teams.
Google has opened Claude access to engineers through Antigravity.
Before this change, Google placed more focus on its own Gemini tools. Access to outside coding tools was more limited.
The new approach gives engineers more choice.
Claude can now work alongside Gemini inside Google's development environment. Google uses usage quotas for Claude access.
This does not mean Google is replacing Gemini.
Instead, Google can let engineers choose the model that fits a specific task.
That can be useful when developers work on large or complex projects.
The main reason is likely engineering productivity.
AI coding has moved beyond simple code suggestions.
Modern AI coding agents can:
Understand large codebases
Plan software changes
Write and edit code
Create tests
Find bugs
Work across files
Use development tools
Handle longer tasks
Google built Antigravity around this type of workflow.
Antigravity lets AI agents work across the editor, terminal, and browser. Agents can plan tasks, run them, and check their work.
Giving engineers access to Claude adds another model to this workflow.
Engineers can then choose the model that works best for a task.
Google and Anthropic are competitors in AI.
Gemini is Google's main AI model family.
Claude is Anthropic's AI model family. It has become especially important for coding and professional work.
Anthropic launched Claude Opus 5 in July 2026. Anthropic says Opus 5 reached the top level on some coding and knowledge-work tests. It also highlights strong results on software engineering tasks.
Google is also improving Gemini.
Google recently launched new Gemini models for agentic work and cybersecurity.
This creates an interesting situation.
Google has its own strong AI models.
Yet Google is also giving engineers access to Claude.
The message is simple:
The best model may depend on the task.
One model may work better for debugging. Another may work better for code review. A third may handle a large software project more effectively.
Antigravity plays a major role in Google's AI coding strategy.
Google introduced Antigravity in November 2025 as an agentic development platform. It lets AI agents work across the editor, terminal, and browser.
Google has since expanded the platform.
Antigravity 2.0 can manage multiple agents and projects. It also supports scheduled tasks and other agent features.
The platform now supports several reasoning models.
Google's current documentation lists Gemini models alongside Claude and GPT-OSS models. Model access depends on the user's plan.
This makes Antigravity more than a traditional coding editor.
It is becoming a place where developers can manage different AI agents and models.
Google is one of the world's largest technology companies.
It also competes directly with Anthropic in AI.
Yet Google is now giving engineers access to Claude.
That is a strong signal for Anthropic.
It shows that Claude has become important enough to earn a place in Google's engineering workflow.
Anthropic also has a strong relationship with Google Cloud.
In April 2026, Anthropic announced a major agreement with Google and Broadcom. The deal covers multiple gigawatts of Google TPU capacity. The new capacity is expected to come online from 2027.
Anthropic also reported more than $30 billion in run-rate revenue at the time. It said more than 1,000 business customers were spending over $1 million per year.
Google has also committed up to $40 billion to Anthropic.
The first investment is $10 billion. Another $30 billion could follow if Anthropic meets certain performance targets.
This creates an unusual relationship.
Google competes with Anthropic. Google also invests in Anthropic and provides computing power to the company. Now, Google engineers can use Claude too.
It would be wrong to say that Google is replacing Gemini.
There is no evidence of that.
Gemini remains central to Google's AI strategy.
Google also continues to build Gemini models for coding and agentic tasks.
At the same time, Antigravity gives users access to several models. Google's current documentation lists Gemini, Claude, and GPT-OSS options.
The better way to understand Google's strategy is model choice.
An engineer can use Gemini for one task.
The same engineer can use Claude for another.
This gives Google a chance to see how different models perform in real engineering work.
That can be more useful than looking at benchmarks alone.
Google's move shows how software development is changing.
AI coding tools are becoming more than simple assistants.
AI agents can now handle larger parts of a development task.
For example, an agent can:
Understand a software request.
Create a plan.
Edit several files.
Run tests.
Find errors.
Make changes.
Check the final result.
This changes the role of the developer.
Developers still make important decisions.
But AI can handle more of the repetitive work.
The model behind the agent also matters.
Different models can produce different results.
That is why Claude vs Gemini is becoming an important comparison for developers.
Developers should watch this trend closely.
Large technology companies are moving toward multi-model AI development.
Developers may use:
Gemini
Claude
GPT models
Open-source models
Specialized coding agents
This creates more flexibility.
A developer does not have to depend on one AI model for every task.
It also creates more competition between AI companies.
Model providers now need to compete on:
Coding quality
Reasoning
Speed
Cost
Context handling
Agent reliability
Tool use
Security
Better competition can give developers better tools.
No clear evidence suggests that.
Google still has a major investment in Gemini.
Claude is better viewed as an additional option.
Google can use Gemini as its core AI platform while giving engineers access to Claude for specific tasks.
This approach also gives engineers more freedom.
If one model struggles with a task, they can try another.
That flexibility could become a normal part of software development.
Google's Claude decision points to a larger trend.
AI companies are no longer competing only on model quality.
They are also competing on developer platforms and AI agents.
The battle now includes:
AI coding agents
Developer platforms
Cloud services
APIs
IDE integrations
Agent frameworks
Enterprise tools
Model selection
Antigravity is part of this shift.
Claude Code is another example.
OpenAI and other companies are also building AI coding agents.
The next stage of AI competition may not be about finding one model that wins every task.
It may be about building the best system around several models.
Google giving engineers access to Claude is an important signal.
It shows that even a major AI company sees value in using rival models.
It also shows that AI coding has become a core engineering tool.
For Google, the goal is better developer productivity.
For Anthropic, the move strengthens Claude's position in professional software development.
For developers, it means more choice.
And for the AI industry, it means the model race is becoming a platform and workflow race.
The most important question may no longer be:
Which AI model is the best?
It may be:
Which AI model is best for this task?
That is the question Google's engineers can now answer with real-world development work.
Why is Google giving engineers access to Claude?
Google is giving engineers another AI coding option. Engineers can use Claude through Antigravity alongside Google's Gemini models. Usage limits apply.
Does Google prefer Claude over Gemini?
No. Gemini remains a major part of Google's AI strategy. Google is adding Claude as another option for engineering work.
Can Google engineers use Claude for coding?
Yes. Current reporting says Google has expanded Claude access to engineers through Antigravity.
What is Google Antigravity?
Google Antigravity is an agentic development platform. It lets developers use AI agents across the editor, terminal, and browser. Agents can plan and execute software tasks.
Why is Claude important for coding?
Claude Opus 5 has shown strong results on coding and knowledge-work evaluations. Anthropic says it is the new state-of-the-art on some of these tests.
What does this mean for developers?
It means developers may increasingly work with multiple AI models. Instead of relying on one model, they can choose the model that fits each task.