Today, we’re introducing [schema]: a harness reaching 99% RHAE with Opus 4.8 + Fable 5 and 95.35% with GPT-5.6 Sol on ARC-AGI-3 Public set. [schema] makes an LLM think like a physicist. 🧵
@naval·Jul 16, 2026·2 sources·neutral
Read articleAI Summary
A new harness called [schema] achieves high performance on the ARC-AGI-3 benchmark by making an LLM think like a physicist.
Related Projects
All Sources
Today, we’re introducing [schema]: a harness reaching 99% RHAE with Opus 4.8 + Fable 5 and 95.35% with GPT-5.6 Sol on ARC-AGI-3 Public set. [schema] makes an LLM think like a physicist. 🧵
@naval
Jul 16, 2026
From our team at Impossible Research - an agent harness that plays games, writes code, reasons like a physicist, and saturates the ARC-AGI-3 benchmark. https://x.com/havenfeng/status/2077770348876247502
@ivanbogatyy
Jul 16, 2026