Antigravity will remove Claude 4.6 models and GPT-OSS-120b on November 18
Google's Antigravity models page says Claude Sonnet 4.6, Claude Opus 4.6 and GPT-OSS-120b will be removed on November 18, 2026. As of November 14 they are still listed. After removal, Free and Google AI Plus keep Gemini 4 Hydrogen and Gemini 3.8 Flash, while Pro, Ultra and Enterprise keep Gemini 4 and Claude 5.5 models.
SessionWatcher editorial · Published · Updated · Source announcement: 2026-10-05
Related guide: How to check your Gemini usage
What is being removed and when
Google's Antigravity models page marks three models with a note that they will be removed on November 18, 2026: Claude Sonnet 4.6 (thinking), Claude Opus 4.6 (thinking) and GPT-OSS-120b. As of November 14, 2026, all three still appear in the model selector and in the plan table. The page covers the Antigravity App, CLI and IDE.
Right now the table lists Claude Sonnet 4.6 and Claude Opus 4.6 for Free and Google AI Plus and for Google AI Pro, but not for Ultra or Enterprise. GPT-OSS-120b is listed for Free and Google AI Plus, Pro and Ultra, but not for Enterprise.
Sources: Models
What each plan keeps after November 18
Free and Google AI Plus: Gemini 4 Hydrogen and Gemini 3.8 Flash. Once the removals happen, the table lists no Claude or GPT model for these plans.
Google AI Pro: Gemini 4 Hydrogen and Gemini 3.8 Flash. Gemini 4 Argon and Orchestrator (Argon) are listed for paid Pro plans or for trials on eligible premium devices. Claude Sonnet 5.5 and Claude Opus 5.5 (thinking) are listed for non-trial Pro subscriptions only. Gemini 4 Carbon and Orchestrator (Carbon) are not listed for Pro.
Google AI Ultra: Gemini 4 Carbon, Gemini 4 Argon, Gemini 4 Hydrogen, both Orchestrator options, Gemini 3.8 Flash, and Claude Sonnet 5.5 and Opus 5.5. Of the three models being removed, Ultra loses only GPT-OSS-120b.
Enterprise: none of the three models being removed is listed for Enterprise, so its list does not change. It keeps Gemini 4 Carbon, Argon and Hydrogen, Gemini 3.8, 3.7 and 3.6 Flash, Gemini 3.1 Pro, and Claude Sonnet 5.5 and Opus 5.5. Orchestrator is not listed for Enterprise.
Sources: Models
What it means for your usage limits
The usage view in the model selector shows two separate meters: one for Gemini models and one for Claude and GPT models. Each meter has its own weekly limit and five-hour limit. All three models being removed use the Claude and GPT meter. If you move that work to a Gemini model, it draws from the Gemini meter instead. On Free and Google AI Plus, the table lists no Claude or GPT model after November 18.
The models page gives no limit sizes. It sends readers to the plans page for reasoning model rate limits. It also says Orchestrator uses fewer tokens than running the planner model alone.
Sources: Models
Corrections
2026-11-05: Updated factual details after a change in the linked official source.
2026-11-14: Updated factual details after a change in the linked official source.
AI assisted reporting, checked against the linked official sources. Source pages checked 2026-11-14. Editorial process and corrections.