Fetching from the wire…
01
02
03
04
05
06
07
08
Research2026-08-02 · source-backed
Tycho formalizes ARC-AGI-3 games as parameterized rendered deterministic Moore machines and has a coding agent build, test, repair, or bypass a free-form executable hypothesis during play. Across all 25 public games under matched budgets, actor-requested delegation scored 88.49 mean RHAE; with that policy GPT-5.6 Sol and Opus 5 both hit 100.00 and completed all 183 levels. The instructive negative: automatic repair after verification failures produced simulators that matched observed transitions far better but reached only 83.07 RHAE. Reproducing dynamics isn't the same as identifying the objective.
Each link below shares sources, entities, or timing with this story.
Opus built by Anthropic / Shared entities / Shared topic / Earlier coverage
Linked by a graph relationship (Opus built by Anthropic); both cover AGI, ARC, GPT, Opus; overlapping topics (agent, arc-agi-3, better).
Claude Code uses Opus / Shared entities / Same source domain / Earlier coverage
Linked by a graph relationship (Claude Code uses Opus); both cover AGI, ARC, GPT, Opus; reported by the same outlet (arxiv.org).
Claude Code uses Opus / Shared entities / Shared topic / Earlier coverage / Tension
Linked by a graph relationship (Claude Code uses Opus); both cover GPT, Opus; overlapping topics (agent, coding).
Claude Code uses Opus / Shared entities / Earlier coverage
Linked by a graph relationship (Claude Code uses Opus); both cover AGI, ARC, Opus; earlier AGI coverage from 2026-07-26.
Claude Code uses Opus / Shared entities / Same source domain / Earlier coverage
Linked by a graph relationship (Claude Code uses Opus); both cover AGI, ARC; reported by the same outlet (arxiv.org).
Claude Code uses Opus / Shared entities / Shared topic / Earlier coverage
Linked by a graph relationship (Claude Code uses Opus); both cover GPT, Opus; overlapping topics (agent, coding).
Linked by a graph relationship (Claude Code uses Opus); both cover GPT, Opus; overlapping topics (agent, coding).
Linked by a graph relationship (Claude Code uses Opus); both cover GPT, Opus; overlapping topics (agent, coding).