Energy-Based Reasoners: 345x less inference compute, 99.9% on Sudoku-Extreme
du_yilun · x · 2026-11-18
A new class of Energy-Based Reasoners (EBRs) proposes an alternative to looped reasoning:
- 345x less inference compute and 2.6x fewer training FLOPs than the best baselines
- 7x–43x more sample-efficient at inference than the best looped baselines
- 99.9% accuracy on Sudoku-Extreme and near 100% on multi-solution benchmarks
The author promises a full thread with details — a notable new direction in reasoning architectures.
More from Research
- 'Physics Hasn't Progressed in 50 Years?' This New PRL Paper on Icy Friction Says Otherwise — CatAstro_Piyush · 2026-11-19
- TMol brings Rosetta-style all-atom molecular energetics into differentiable PyTorch — gottapatchemall · 2026-11-19
- ELLIS Tübingen safety team releases horizon-labs-agent-traces dataset on Hugging Face — NielsRogge · 2026-11-19
- Noam Brown confirms the math model claim; speculation points to an astra math fine-tune — haider1 · 2026-11-19
- NCME talk: generative psychometrics validates LLM-generated test items before human data — GolinoHudson · 2026-11-19
- Researchers say peer review has hit a breaking point as AI conference scale strains ICLR — LucaAmb · 2026-11-19