GitHub is making Google's new AI model generally available in its Copilot development assistant. Eligible paid plan users can choose their preferred model in editors and apps, with the new option rolling out gradually.
1. A new model for coding workflows that chain tools
GitHub announced general availability for Gemini 4 Neon on November 18, 2026. The model targets agentic coding, where AI uses tools to carry out coding tasks. It also targets multistep tool use and knowledge work.
Developers select the new model through Copilot's existing model picker. The key change is the ability to switch coding assistance models within the same workspace, rather than moving to a separate service.
2. Google emphasizes similar quality at a lower cost per task
According to Google, Gemini 4 Neon offers quality close to Gemini 4 Carbon at a lower token cost per task. The comparison focuses on reducing cost per task while maintaining similar quality, rather than greatly improving quality.
Google said it starts responding sooner and makes fewer unrequested changes in coding tasks than the comparison model. The former concerns waiting time for answers, while the latter concerns the scope of requested edits.
Google also said the model is less likely to make unsupported claims about code, tools, and test results. This describes improvements to both output quality and how the model explains its work. The announcement gives no figures or measurement conditions, so the scale of improvement cannot be calculated.
3. Available through model pickers, from editors to mobile
Eligible users are on Copilot Pro+, Max, Business, and Enterprise plans. In supported environments, they can select the new model from the model picker. The rollout is gradual, and GitHub advises checking again later if it is not yet listed.
Supported development tools include Visual Studio Code and Visual Studio. The model can also be selected in JetBrains IDEs, Xcode, and Eclipse. The model picker in the user's usual editor is the starting point.
Availability also covers Copilot CLI, the command-line tool, and GitHub Copilot coding agent, which performs coding tasks. Users can select it in the GitHub Copilot app and on GitHub's website. GitHub Mobile includes both iOS and Android apps.
4. Administrator model policies apply to organization users
Business and Enterprise administrators can manage access through model policies in Copilot settings. This lets administrators control model use across their organizations.
The new model is enabled automatically when the default model enablement setting is on. Automatic enablement does not apply if an administrator disables the overall default setting or explicitly disables Gemini 4 Neon.
5. Weigh per-task cost benefits alongside usage-based billing
The expected benefit is Google's stated combination of similar quality and lower token costs per task. Gemini 4 Neon uses usage-based billing at the model provider's list prices. Model selection therefore involves weighing expected usage and costs alongside quality and response characteristics.
Pricing is available in the Copilot model and pricing documentation linked by GitHub. Comparing the announcement's per-task cost claims with the prices used for actual billing is a starting point for adoption decisions.
Gemini 4 Neon joins Copilot's existing model selection workflow, letting users weigh gradual availability and usage-based costs when choosing it.
