Will GPT-5.2 Achieve an AA-Omniscience hallucination rate below 40%?
4
Ṁ2347Jan 31
14%
chance
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
This question is managed and resolved by Manifold.
Get
1,000and
3.00
1,0003.00Sort by:
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.
Related questions
Related questions
Will I be impressed by GPT-5?
19% chance
Will GPT-5 get the Monty Fall problem correct?
95% chance
Before 2027, will a frontier AI model achieve an AA-Omniscience hallucination rate below 5%?
43% chance
Will GPT-5 make Manifold think AI-caused extinction is more likely?
12% chance
Will GPT-5 have Atari skills?
3% chance
Will GPT-5 destroy the world?
1% chance
Will hallucinations (made up facts) created by LLMs go below 1% on specific corpora before 2025?
38% chance
Will GPT-5 be capable of achieving superhuman performance in at least one exam that is typically taken by humans?
91% chance
Will GPT-5 make Manifold think very near-term AGI is more likely?
11% chance
Before 2027, will OpenAI release a frontier model with a 5:1 or better abstention to hallucination ratio on SimpleQA?
51% chance