Will openAI have the most accurate LLM across most benchmarks by EOY 2024?

Benchmark Open-LLM or whatever is sota to compare LLMs by EOY in 2024: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard

Get Ṁ600 play money
Sort by:

@JanCarbonell This market has a closing date of 2023-12-31 but the title indicates the closing date should be 2024-12-31. I am changing the closing date to match the title. If you don't think this is right, write in!

@JanCarbonell it looks like that dashboard is only for open source LLMs. Is this question only about open-source LLMs, and therefore OpenAI can only win it if they release an open-source model (which I think I have before but it's an old and not very capable one)?

@chrisjbillington Chris, this is a good and important question. With 17 participants already, there has been some action, but the market looks unsure at the moment.

I would recommend any user who is interested in this question create a new version with more clarity in the criteria and post it here. If the creator does not show up to clarify it, this market might (or might not) end up as N/A. If you'd rather play in a better-defined market, make one!

predicts NO

@chrisjbillington Great question, I did not think about that when setting the market. What do you think would be the best way to benchmark both OSS and closed source LLMs?

More related questions