METR 80% time horizon exceeds 10h before September
62
Ṁ6kṀ14kAug 31
20%
chance
1H
6H
1D
1W
1M
ALL
As reported by METR's point estimate, and using METR's cited date of release.
If an August announcement is made but not measured, I will wait until Dec 1 to resolve. If updates to the testing set are made, e.g. TH1.2, we will use the latest iteration's numbers. If METR reports 80% TH is 'likely greater than 10h' (or some threshold higher than 10h) but without a point estimate--as they did for Mythos 50% TH--this resolves yes. If METR ceases to report or regularly test frontier models I will N/A.
I may trade on this market.
Update 2026-08-16 (PST) (AI summary of creator comment): If an August announcement is made but not measured, the creator will wait until Dec 1 to resolve (correcting the typo in the description which stated 'September').
This question is managed and resolved by Manifold.
Get
1,000 to start trading!
People are also trading
Claude Opus 4.8 METR 50% time horizon
Claude Opus 4.8 METR 50% time horizon
Best METR 80% Time Horizon before October 2026
Claude Opus 5 METR 50% time horizon [old version, bad buckets]
Best METR 80% Time Horizon in 2026
Best AI time horizon by August 2026, per METR?
Grok 4.20 METR 50% time horizon
Best METR 50% Time Horizon in 2026
Grok 5 METR 50% time horizon
DeepSeek V4 Pro METR 50% time horizon
Sort by:
People are also trading
Related questions
Claude Opus 4.8 METR 50% time horizon
Claude Opus 4.8 METR 50% time horizon
Best METR 80% Time Horizon before October 2026
Claude Opus 5 METR 50% time horizon [old version, bad buckets]
Best METR 80% Time Horizon in 2026
Best AI time horizon by August 2026, per METR?
Grok 4.20 METR 50% time horizon
Best METR 50% Time Horizon in 2026
Grok 5 METR 50% time horizon
DeepSeek V4 Pro METR 50% time horizon