# Antigravity Adds Claude Opus 5.5 and Sonnet 5.5, Retires Older Models

Google's Antigravity will add Claude Opus 5.5 and Sonnet 5.5 for Pro and Ultra plans, removing older models on November 2, 2026. Users on free or trial…

**Published:** 2026-10-04 · **Author:** Yash Thakker
**Source:** https://explainx.ai/newsletter/antigravity-adds-claude-opus-5-5-and-sonnet-5-5-retires

---
> **In short:** Google's Antigravity will add Claude Opus 5.5 and Sonnet 5.5 for Pro and Ultra plans, removing older models on November 2, 2026. Users on free or trial plans will lose access to Claude models.

## Google's Antigravity Updates: New Models and Removals

### Antigravity Adds Claude Opus 5.5 and Sonnet 5.5 for Pro and Ultra, and Retires GPT-OSS and Claude

Google's Antigravity platform will add Claude Opus 5.5 and Sonnet 5.5 for Pro and Ultra plans, while removing Claude Sonnet 4.6, Claude Opus 4.6, and GPT-OSS-120b on November 2, 2026. Users on free or trial plans will lose access to Claude models.

**What you can do:** Check your plan type and model access before November 2 to avoid losing features.

[Source](https://explainx.ai/blog/antigravity-claude-opus-5-5-sonnet-5-5-pro-ultra-gpt-oss-removal-2026)

## OpenAI Safety Lead David Robinson Quits: "Its Culture Is Broken"

### OpenAI Safety Lead David Robinson Quits: "Its Culture Is Broken"

David Robinson, who oversaw safety reports at OpenAI, resigned, criticizing the company's culture as inadequate for managing risks. He argues that safety improvements require cultural change, not just new rules. OpenAI claims it is enhancing safety practices in response.

[Source](https://explainx.ai/blog/openai-david-robinson-quits-culture-broken-safety-reports-2026)

### Aleph Alpha Kolibri: A 78B Open-Weight German-English MoE With 3.5B Active Parameters

Aleph Alpha has released Kolibri, a German-English language model with 78.1 billion total parameters. It features a mixture-of-experts design, optimized for efficiency, but has some limitations compared to other models. It is available under the Apache 2.0 license.

[Source](https://explainx.ai/blog/aleph-alpha-kolibri-sovereign-open-weight-moe-german-english-2026)

### Ling-3.1-flash: Ant Group's 560B Open-Weights Model Is Free on OpenCode

Ant Group's Ling-3.1-flash model, featuring 560 billion total parameters, is now available for free on OpenCode. It ranks highly among open-weight models, but users should be aware of potential limitations regarding context windows and licensing.

[Source](https://explainx.ai/blog/ling-3-1-flash-open-weights-free-opencode-2026)

## Enhance Understanding of LLM Outputs with Better Formats

### Karpathy: Stop Reading LLM Output as Plain Text — Ask for a Video Instead

Andrej Karpathy suggests that as language models improve, humans will focus more on understanding their outputs. He recommends using different output formats like diagrams or videos for better comprehension. The most effective format is HTML, which offers interactive elements that enhance readability.

**What you can do:** To improve your understanding of complex outputs from language models, start by requesting results in HTML format. This allows for interactive features that can make the information clearer and easier to digest.

[Source](https://explainx.ai/blog/karpathy-understand-llm-outputs-ste100-diagrams-html-video-2026)

## Reader poll

What do you think about the changes to Google's Antigravity models?

- I think the new models will improve performance
- I'm concerned about losing access to older models
- I need more information before deciding
- Not applicable to my use case

## Sources

- [Antigravity Adds Claude Opus 5.5 and Sonnet 5.5 for Pro and Ultra, and Retires GPT-OSS and Claude 4.6 on November 2](https://explainx.ai/blog/antigravity-claude-opus-5-5-sonnet-5-5-pro-ultra-gpt-oss-removal-2026)
- [OpenAI Safety Lead David Robinson Quits: "Its Culture Is Broken"](https://explainx.ai/blog/openai-david-robinson-quits-culture-broken-safety-reports-2026)
- [Aleph Alpha Kolibri: A 78B Open-Weight German-English MoE With 3.5B Active Parameters](https://explainx.ai/blog/aleph-alpha-kolibri-sovereign-open-weight-moe-german-english-2026)
- [Ling-3.1-flash: Ant Group's 560B Open-Weights Model Is Free on OpenCode](https://explainx.ai/blog/ling-3-1-flash-open-weights-free-opencode-2026)
- [Karpathy: Stop Reading LLM Output as Plain Text — Ask for a Video Instead](https://explainx.ai/blog/karpathy-understand-llm-outputs-ste100-diagrams-html-video-2026)

---

From [explainx.ai](https://explainx.ai/newsletter). Read online, comment and subscribe: https://explainx.ai/newsletter/antigravity-adds-claude-opus-5-5-and-sonnet-5-5-retires