Antigravity 2.0

Overview

Antigravity 2.0 is Google’s coding agent and the model behind the agentic-IDE surface that the May 22 HN “bait and switch” critique was originally aimed at. In May 2026 the model took #1 on Modelrift’s OpenSCAD architectural-3D LLM benchmark — a niche but practitioner-built eval — with sustained HN discussion (369 pts / 146 cmts).

Timeline

• 2026-05-23-AI-Digest — Tops Modelrift’s OpenSCAD architectural-3D LLM benchmark at #1; community-built eval, but the leader isn’t a typical general-coding favorite, useful as a model-specific-strengths data point to hold for cross-reference when other structured-spatial-output evals appear. HN frontpage discussion 369 pts / 146 cmts.

Key Developments

1. #1 on Modelrift OpenSCAD Architectural-3D Benchmark: Community-built benchmark; niche eval, but a non-default general-coding leader landing top-of-list on a structured-spatial-output task is the kind of model-specific data point worth tracking against future structured-output evals.

Related

See also: Google, MOC - Agentic Coding, MOC - Developer Tools.