METR's AI time-horizon chart hasn't added a model since May 8
mhmazur · x · 2026-10-12
haider1 asks why METR's famous AI time-horizon chart — measuring how long a task a model can complete with 80% success — suddenly stopped updating, with the last model added on May 8. Given how much frontier models have improved since, he's curious how far Opus 5.5 and Astra could push the horizon, and hopes METR resumes the evaluations. A reply jokes "maybe they're still running."
Related event: METR's AI Time Horizon Benchmark Stalls for Nearly Six Months(2 posts)→
More from Models
- ThursdAI: OpenAI drops 722 math papers, Haiku 5.5 hits 10 cents, Meta's new agent protocol — altryne · 2026-10-12
- Mercury Decide nails 669 clinical decisions for 1 cent, ties #2 at 94.2% cheapest — MaziyarPanahi · 2026-10-12
- Meta employees hint at a major new model launch, rumored as early as this week — kimmonismus · 2026-10-12
- LEAN proofs can be 'correct but alien' or exploit kernel bugs, researcher warns — gklambauer · 2026-10-12
- Four ways AI-generated LEAN proofs can go wrong, from formalization drift to kernel bugs — gklambauer · 2026-10-12
- Opus 5.5 Stuns With 3D Car Models It Can Even Split Into Parts — iamfakhrealam · 2026-10-12