
How much cheaper to use will o3-equivalent or better models get before 2026?
26
1kṀ4860Dec 31
95%
≥2x
83%
≥5x
71%
≥10x
55%
≥30x
39%
≥100x
Any model with publicly known benchmark scores and inference costs goes, not just OpenAI's o series.
I will consider a model to be "o3-equivalent or better" if it scores ≥25% on FrontierMath (o3 scored 25.2%) and performs similarly on other benchmarks.
(Note that o3's exact inference costs in the configuration used for benchmarking are currently unknown IIUC, though this market description will be updated with exact figures if they become public. This market can still resolve even without exact figures if e.g. OpenAI announce an o4 that's "10x cheaper" for roughly the same performance.)
This question is managed and resolved by Manifold.
Get
1,000 to start trading!
People are also trading
Related questions
Will o3.5 be released before 2026?
4% chance
Which of the following "colors" of hydrogen will be the cheapest by 2030? [details in description]
Will there be a reasoning model more powerful than o1-preview, and cheaper and >10x faster than o1-mini, by Nov 12 2025?
84% chance
Will OpenAI launch a model even more expensive than o1-pro in 2025?
36% chance
Will o4 be released before 2026?
72% chance
Will producing deuterium be cheaper in 2027 than 2025?
62% chance
What will the inference cost of the best publicly available LM be in 2030?
Will a single model have all the upsides o1-style RL with none of the downsides at 2027?
58% chance
During what year will the cost of CO2 sequestration per tonne drop by 10x from 2023 prices?
Will o5 be released before 2026?
28% chance