Google's Antigravity SDK adds local model support: run agents offline with Gemma 4 and LiteRT
GlennCameronjr · x · 2026-09-24
Google officially announced that the Antigravity SDK now supports local model execution, debuting with Gemma 4 26B A4B on Google AI Edge's LiteRT runtime, enabling fully offline agentic workflows.
Key selling points of local execution:
- Zero token costs — no API fees or rate limits
- Privacy — code and requests stay entirely on-device, ideal for compliance-restricted environments
- Offline resiliency, plus hybrid workflows combining local and cloud models
- OpenAI-compatible endpoint support, so Ollama, llama.cpp, and vLLM can serve Gemma
Getting started: pip install google-antigravity litert-lm, then download the Gemma 4 26B A4B model from Hugging Face. Google recommends a machine with >24GB VRAM or unified memory.
Related event: Google Antigravity SDK Adds Local AI Model Support with Gemma 4(9 posts)→
More from coding & agent
- As models improve, writing good evals gets harder and more important — amankhan · 2026-11-12
- Omarchy 4.6 lets you share local AI models across your tailnet, says DHH — alexcovo_eth · 2026-11-12
- Coding harnesses hand shell access to inference providers — 4-step gateway mitigation — shensi · 2026-11-12
- Slack's David Hoang: agents beat people on coordination, proposes 'boundary agents' — davidhoang · 2026-11-12
- CS student mourns coding flow states he'll never experience in the AI era — chrissyykat · 2026-11-12
- AI Engineer Code Summit features its first all-female keynote lineup — altryne · 2026-11-12