feat: launch Gemini 3 in Gemini CLI 🚀🚀🚀#13279
Conversation
* basic changes * remove logic in Banner component * works with placeholder text; wait for banner text feature * works now * cleanup * lint * finishing touches * address bots comment * use colors in figma and address nit * add tests * add padding to the right * fix test
* checkpoint * some fixes * various fixes * rename to previewModel * checkpont * retry only once * address issues * change model name * add 500 fix
…30) * Show the preview model banner only if the previewFeature is disabled * refactor and add tests
* fix(core): restrict retry logic to gemini-2 models * fix
* wait for the banner data * fix test * cancel the fetch on unmount
Moves EDITOR_DISPLAY_NAMES and adds getEditorDisplayName helper to packages/core/src/utils/editor.ts. Updates CLI components to use these centralized definitions, ensuring consistent editor naming across the application.
* Enable include_thoughts for more models * Update tests * Add model
* Move previewFeatures to the top of the settings * update snapshot for settings dialog test * use gradient color for the banner and update the modelDialog UI * fix test * regenerate schema:settings * update settings doc
* fix(core): Restrict retry logic to Gemini 2.x models (#28) * fix(core): restrict retry logic to gemini-2 models * fix * add preview feature availability * change flag name and allow undefined values * address nit and prevent integ test from failing * nits * address comment on model override * remove comment * make previewFeatures for all users non Google auth and stop setting model * update tests and logic * test * add back --------- Co-authored-by: Sandy Tao <sandytao520@icloud.com> Co-authored-by: gemini-cli-robot <gemini-cli-robot@google.com>
* complete * fix tests
Introduces a `PersistentState` utility to track local user state across sessions. Updates `AppHeader` to hide the default welcome banner after it has been shown 5 times to reduce visual noise for experienced users. - Added `packages/cli/src/utils/persistentState.ts` for managing `state.json`. - Updated `AppHeader.tsx` to increment and check `defaultBannerShownCount`. - Added unit tests for banner display logic. - Minor: Updated model name in `geminiChat.test.ts`.
* Quota/capacity alert message udpates * Fallback messags UI updates * fix text and update the gemini-3 model name * remove gemini 2 mode checking to enable capturing error thrown to show the error to the user * Revert "remove gemini 2 mode checking to enable capturing error thrown to show the error to the user" This reverts commit c28042c1a62ace2ef53923623bf4ad8944f25844.
* update to ids * rm logs
… not undefined. (#48)
* complete * add comment * add working test * update text
* unescaped newline that's escaped * Add test
* Add cli/settings.md documentation * Add get started page. * Add to indexes and sidebar. * Specify models. * Clarification. * Fix typo. * Index links. * Add new content for gemini-3.md * Updates. * Formatting. * Clarify model usage on model.md * Update sidebar.
Summary of ChangesHello @skeshive, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request significantly upgrades the Gemini CLI by integrating Gemini 3 Pro, Google's most intelligent model, for users with access. It introduces a new 'Preview features' setting to enable this functionality and overhauls the model selection experience to reflect these advanced options. The changes also include robust improvements to error handling for model-related issues, offering users clearer guidance and choices during quota or capacity limitations. Furthermore, IDE integration has been broadened to support Antigravity, enhancing the CLI's usability across more development environments. Accompanying these functional updates, extensive documentation has been added and revised to ensure users can easily navigate and utilize the new features and settings. Highlights
Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here. You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension. Footnotes
|
|
Size Change: +20 kB (+0.1%) Total Size: 20.6 MB
ℹ️ View Unchanged
|
skeshive
changed the title
3️⃣
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces support for Gemini 3 Pro as a preview feature. The changes are comprehensive, spanning documentation, core logic for model selection and error handling, UI components for user notification and configuration, and extensive testing. The documentation updates are excellent, providing clear instructions for users. The core logic for handling the preview model, including fallback and bypass modes for different failure scenarios, is robust and well-implemented. The UI has been enhanced with a new banner and improved dialogs for model selection and error handling, which will significantly improve the user experience. The testing strategy is thorough, with new tests covering the complex new logic. Overall, this is a high-quality submission, and I have not identified any issues of high or critical severity.
https://goo.gle/enable-preview-features gives me a 404 page... 
|
15 participants
Gemini 3 Pro, our most intelligent model, now available in your terminal!
We are rolling out access gradually to ensure the experience remains fast and reliable.
You can also track our rollout progress by following this GitHub discussion.
If you have access, you can enable Gemini 3 by running
/settings, then toggle “Preview features” to “true.” Gemini CLI will now default to Gemini 3 Pro.Enabling.Gemini.3.mp4
Read our 5 things to try with Gemini 3 Pro in Gemini CLI blog to get some inspiration for what is capable.
Golden.Gate.Simulation.mp4
Contributors: @abhipatel12, @adamfweidman, @Adib234, @aishaneeshah, @galz10, @jacob314, @joshualitt, @SandyTao520, @sehoon38, @skeshive, @NTaylorMullen