Will GPT-5.2 Achieve an AA-Omniscience hallucination rate below 40%? | Manifold

Will GPT-5.2 Achieve an AA-Omniscience hallucination rate below 40%?

4

1kṀ2347

Jan 31

14%

chance

8

1H

6H

1D

1W

1M

ALL

Resolves according to: https://artificialanalysis.ai/?omniscience=omniscience-hallucination-rate

Resolves if OpenAI releases a frontier model that is clearly a successor to GPT-5.1, even if it isn't called GPT-5.2

Technical AI Timelines

Get

1,000

to start trading!

Sort by:

bought Ṁ600 NO

@creator resolve no got 78% https://artificialanalysis.ai/evaluations/omniscience

Only one model (Claude Opus 4.1) has gotten that good so far. HOWEVER, out of OpenAI models, the scores have been steadily improving (GPT-5.1 > GPT-5-mini > GPT-5 > gpt-oss-120B > gpt-oss-20B > GPT-4.1; though do note that GPT-5-mini, GPT-5, and the gpt-oss models were released very close to each other), so it seems plausible that we'd reach it someday.

People are also trading

Will I be impressed by GPT-5?

Will GPT-5 get the Monty Fall problem correct?

Before 2027, will a frontier AI model achieve an AA-Omniscience hallucination rate below 5%?

Will GPT-5 make Manifold think AI-caused extinction is more likely?

Will GPT-5 have Atari skills?

Will GPT-5 destroy the world?

Will hallucinations (made up facts) created by LLMs go below 1% on specific corpora before 2025?

Will GPT-5 be capable of achieving superhuman performance in at least one exam that is typically taken by humans?

Will GPT-5 make Manifold think very near-term AGI is more likely?

Before 2027, will OpenAI release a frontier model with a 5:1 or better abstention to hallucination ratio on SimpleQA?

Related questions

Will I be impressed by GPT-5?

Will GPT-5 get the Monty Fall problem correct?

Before 2027, will a frontier AI model achieve an AA-Omniscience hallucination rate below 5%?

Will GPT-5 make Manifold think AI-caused extinction is more likely?

Will GPT-5 have Atari skills?

Will GPT-5 destroy the world?

Will hallucinations (made up facts) created by LLMs go below 1% on specific corpora before 2025?

Will GPT-5 be capable of achieving superhuman performance in at least one exam that is typically taken by humans?

Will GPT-5 make Manifold think very near-term AGI is more likely?

Before 2027, will OpenAI release a frontier model with a 5:1 or better abstention to hallucination ratio on SimpleQA?

© Manifold Markets, Inc.•Terms•Privacy