Skip to main content
MANIFOLD
Will a model achieve a METR 50% time-horizon of 4+ hours by the end of 2025?
29
Ṁ100Ṁ3.2k
resolved Dec 20
Resolved
YES

METR's 50%-time horizon doubling time trendline from more recent models indicates that a model released towards the end of 2025 will likely have a time horizon of over 4 hours. Will this come true?

See links for more details:
https://www.alignmentforum.org/posts/SuvWoLaGiNjPDcA7d/metr-s-evaluation-of-gpt-5
https://www.lesswrong.com/posts/deesrjitvXM4xYGZd/metr-measuring-ai-ability-to-complete-long-tasks

Market context
AI
Technical AI Timelines
AI model releases
AI Benchmarks
METR
Get
Ṁ1,000
to start trading!

🏅 Top traders

#TraderTotal profit
1Ṁ575
2Ṁ286
3Ṁ147
4Ṁ84
5Ṁ76

People are also trading

Sort by:

Opus 4.5 seems to have landed close to the faster trendline!

https://xcancel.com/METR_Evals/status/2002203627377574113#m

@Jasonb I am not so sure yet that it is close to the faster trend line. If you look at the new graph posted by METR they say a doubling of 196 days. 196 days comes to 6.4 months so closer to the original 7 month doubling estimate.

bought Ṁ931 YES

@creator Claude Opus just topped 4 hours!

@MRME incredible

@Bayesian I'm in shock. I thought the curve was going to slow down. I have no words.

https://x.com/SamuelAlbanie/status/1991301408876835285/photo/1

People are also trading