Will a “good” (fine-tuned) opensource model utilizing chain-of-thought reasoning like o1 be released by EOY 2024?
Plus
26
Ṁ4804resolved Dec 20
Resolved
YES1D
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 Release a Reasoning model (a la o1) before OpenAI releases o3 for general users.
52% chance
Will there be a reasoning model more powerful than o1-preview, and cheaper and >10x faster than o1-mini, by Nov 12 2025?
51% chance
Will a Chinese AI developer announce a model rivaling o3 performance by February 2025?
20% chance
Will the state-of-the-art AI model use latent space to reason by 2026?
47% chance
Will OpenAI release a model which performs chain-of-thought on image tokens / visual scratchpad before 2026?
74% chance
Will OpenAI release a model which generates images using reasoning / inference-time scaling before 2026?
63% chance
Will it be public knowledge by EOY 2025 that a major AI lab believed to have created AGI internally before October 2023?
8% chance
Will OpenAI release o2 (or o3) before 2026?
98% chance
Will there be a paper published by 2027 that says reasoning models are flawed?
76% chance
Will Algorithmica receive more content before EOY 2025?
52% chance