Will a “good” (fine-tuned) opensource model utilizing chain-of-thought reasoning like o1 be released by EOY 2024?
Plus
25
Ṁ4754Dec 31
90%
chance
1D
1W
1M
ALL
The model should be able to decide how long it needs to think based on the complexity of the problem. Ideally it should be ranked higher on LMSYS than the “normal” model but this is not a strict requirement. I imagine OLAMA would update their UI to support something like that but also would be ok if someone else does it. If model isn’t ranked higher on LMSYS then the market would resolve based on my experience playing around with it. In general, it has to be significantly better than just using the normal version. Ameaturish attempts will be dismissed. I won’t be betting on this market.
This question is managed and resolved by Manifold.
Get
1,000
and3.00
Related questions
Related questions
Will Anthropic, Google, xAI or Meta release a model that thinks before it responds like o1 from OpenAI by EOY 2024?
61% chance
Will Anthropic release a model that thinks before it responds like o1 from OpenAI by EOY 2024?
23% chance
Which of these companies will release a model that thinks before it responds like O1 from OpenAI by EOY 2024?
Will anyone be able to get OpenAI’s new model o1 to leak its system message by EOY 2024?
28% chance
When will OpenAI release o1?
Will an open source model beat GPT-4 in 2024?
37% chance
Will there be a reasoning model more powerful than o1-preview, and cheaper and >10x faster than o1-mini, by Nov 12 2025?
74% chance
Will there be an open source video model roughly as good as OpenAI Sora by the end of 2024?
26% chance
Will Meta release an open source language model that outperforms GPT-4 by the end of 2024
67% chance
Will OpenAI release o2 before 2026?
79% chance