New paper measures frontier models’ no-CoT task horizons
scaling01 · x · 2026-07-20
A new paper, Think Fast: Estimating No-CoT Task-Completion Time Horizons of Frontier AI Models, looks at how long frontier models can keep completing tasks without chain-of-thought.
The post highlights METR’s observation that model time horizons have been doubling every few months, and uses that as the motivation for estimating no-CoT task-completion horizons more directly.
More from Research
- HarnessMM cuts agent tokens per task by up to 88% while raising success rates — omarsar0 · 2026-10-26
- Ben Recht: The world needs a truly open language model, and it's easier to build than you think — beenwrekt · 2026-10-26
- New Blog Breaks Black-Box Distillation Into Five Practical Steps — chrisrohlf · 2026-10-26
- DetectorScope turns black-box AI text detectors into auditable ones with sparse autoencoders — _akhaliq · 2026-10-26
- OPST paper shows LLMs self-improve with zero human data, guided only by topic descriptions — _akhaliq · 2026-10-26
- Tencent's HiSTEVE 1.0: hierarchical VLA agent plays Minecraft with native keyboard and mouse, hits 85% SOTA — _akhaliq · 2026-10-26