Ex-DeepMind Researchers' Agent Hits 75.5% on ARC-AGI, Invents New Methods
hyperparticle · x · 2026-07-22
Two ML researchers with 21 years of experience (ex-Google, DeepMind, Luma AI) are hiring Founding Scientists for their startup Rekursiv.ai.
They demonstrated impressive benchmark results from their autonomous agent team:
- ARC-AGI-1: Score jumped from 44.9% to 75.5% using solutions the agent invented itself.
- ARC-AGI-2: Successfully transferred the recipe.
- Sudoku: Invented a new method (Hypothesis-Pinning Search), hitting 100% on Sudoku-Extreme by writing 125k+ lines of code and running 684 experiments.
Related event: Rekursiv AI Agents Achieve Major Breakthroughs on ARC-AGI(2 posts)→
More from coding & agent
- Dev turns DJI Mavic into local AI security drone built with Opus 5.5 — alexcovo_eth · 2026-10-27
- Microsoft Research: ForkPilot cuts long-horizon agent compute by up to 59% — EliasEskin · 2026-10-27
- Cursor launches Agent SDK: agents as code with Markdown prompts and TypeScript tools — soleio · 2026-10-27
- Vibecheck, the 'usable formal methods' hackathon, sells out two floors in SF — spikedoanz · 2026-10-27
- How do you verify AI agent outputs beyond unit tests and LLM-as-judge? — Practical_Green5775 · 2026-10-27
- AI agents may soon become the only way to run code, not just write it — IgorCarron · 2026-10-27