TuF Labs tops ARC-AGI-2 leaderboard: general approaches now beat domain-specific tricks
rbhar90 · x · 2026-11-10
Dries Smit's team at TuF Labs wrapped up ARC Prize 2026 with a first-place finish on ARC-AGI-2 and second place on ARC-AGI-3 (public leaderboards).
Their key takeaway: general approaches now beat domain-specific tricks on ARC benchmarks — a notable shift for the competition. Full writeups and code are public.
More from Models
- Fable 5.2 system prompt leaked: 420,000+ characters, full text on GitHub — morgymcg · 2026-11-10
- ChatGPT remote control on Android plagued by failures, users say Claude works better — koltregaskes · 2026-11-10
- Screenshot shows Gemini explaining 'load shedding' of users to cheaper models — Top-Sink-1315 · 2026-11-10
- Fable 5.2 Robot Control Demo Sped Up 60x, Sparking Pushback on "Robot Lab Is Dead" Hype — kaixhin · 2026-11-10
- Seroter's daily reading list: new TPUs, Claude 5.2 price cut, and the future of the PM role — rseroter · 2026-11-10
- Toby Ord revisits METR capability trends: full-data fit yields 1.6x acceleration, not 1.4x — tobyordoxford · 2026-11-10