Users discuss METR benchmark placement for Opus 5, estimating ~20 hours
Anxious-Yoghurt-9207 · reddit · 2026-08-19
A Reddit post asks where current models stand on METR's time horizon score, suggesting Opus 5 might reach 20 hours, seeking updates since benchmarks stalled in May.
More from Models
- Inkling-Small hits Agent Arena: matches Inkling performance at half the price — simonguozirui · 2026-08-19
- LFM2.5 Models and Embedding Now Available on OpenRouter — maximelabonne · 2026-08-19
- Will we run 30B+ parameter models fast on small GPUs in the future? — absurdother · 2026-08-19
- Grok 4.6 hands-on: fast enough to kill context switching, wins over a Codex subscriber — elonmusk · 2026-08-19
- Pangram Clarified: Not a Measure of Writing Quality — TuhinChakr · 2026-08-19
- Zhipu's Tang Jie on Scaling Laws: Inference and MoE Shift Optima — jietang · 2026-08-19