GPT 4.5 "Intelligence" via artificialanalysis.ai
19
100Ṁ1098Jul 1
4%
<50 (underwhelming)
86%
50-54 (at or below o1-mini)
6%
55-59 (worse than Deepseek R1)
1.9%
60-64 (matches SoTA reasoning models)
1.8%
65+ (new frontier)

In May 2024, Orion (GPT-4.5, assumed to be GPT-5 at the time) was promised to be a Whale by Microsoft.

Will it surpass the current whale (Deepseek R1)?
Source: https://artificialanalysis.ai (subject to some discretion in case the methodology changes)
This question is managed and resolved by Manifold.
Get
1,000 to start trading!
Sort by:
bought Ṁ1 YES

Assuming it's not a reasoning model is 60 a fair upper limit?
Or does OpenAI have something up their sleeves to one up Grok 3 (~53 w/o reasoning, ~66 w/ reasoning)
Since there's over 50% for this to resolve to the highest option I added more
https://manifold.markets/ChinmayTheMathGuy/gpt-45-intelligence-via-artificiala-u8ZOQuAhOZ
Related questions
Related questions
GPT 4.5 Intelligence (higher options)
Will GPT-5 like to delve?
36% chance
Will GPT-5 have an estimated IQ higher than 101?
79% chance
Is GPT4 sentient?
12% chance
Will GPT-4 be legally considered AGI?
7% chance
Will GPT-4.5 score at least 100 in an IQ test?
65% chance
What will be true about GPT-5?
Could GPT-4 recursively self-improve to AGI with the right cognitive architecture? [@Altimor, twitter]
9% chance
To what extent is GPT-4 AGI?
POLL
Will LLMs such as GPT-4 be seen as at most just a part of the solution to AGI? (Gary Marcus GPT-4 prediction #7)
91% chance