Will Gemini Ultra achieve an Elo of 1300 or higher on the Chatbot Arena before May 1, 2024?
34
716
650
resolved May 1
Resolved
NO
  • Gemini Ultra was released on Feb 8, 2024

  • As of Feb 9, 2024, it is expected to achieve a higher Elo than GPT-4 according to this market

    • As of Feb 9, 2024, GPT-4 has an Elo of 1253

Resolution criteria:

  • Resolves YES if some version of Gemini Ultra achieves an Elo of 1300 or higher on Chatbot Arena at any point before May 1, 2024

  • Resolves NO if Gemini Ultra does not reach an Elo of 1300, or if it never appears on the leaderboard before the deadline

  • Resolves N/A if the leaderboard is no longer publicly accessible at the time of closing

Get Ṁ600 play money

🏅 Top traders

#NameTotal profit
1Ṁ351
2Ṁ256
3Ṁ209
4Ṁ27
5Ṁ25
Sort by:

Inactive creator, Gemini Ultra has not appeared on the leaderboard, modresolving NO.

@chrisjbillington I was keeping an eye on this, but thanks!

By gemini ultra, you are referring to the one powered by search engine, right? Like the bard(Pro)

opened a Ṁ400 NO at 16% order

@Sss19971997 I am up to betting more, limit order expire in a week

@Sss19971997 By "Gemini Ultra", I am referring to the model that powers Gemini Advanced. You might be thinking of another model, Gemini Pro, that powers Gemini. Gemini Pro shows up on the leaderboard as "Bard (Gemini Pro)", "Gemini Pro (Dev API)", and "Gemini Pro". None of these count as "Gemini Ultra".

@howahlah So Gemini Advanced does not count?

opened a Ṁ100 NO at 14% order

@howahlah When I mentioned Gemini Pro, I was making an analogy. The Bard(Pro) is not an LLM but a search-engine augmented chatbot. So I was asking if Gemini Advanced (Gemini Ultra augmented with search) shows up on the leader board, does it count as Gemini Ultra?

@Sss19971997 Yes, I would consider Gemini Advanced "some version of Gemini Ultra". This is analogous to Gemini Pro showing up as "Bard (Gemini Pro)".

opened a Ṁ150 YES at 16% order

@howahlah Then 1300 seems easy

@Sss19971997 sure, feel free to bet up. Bear in mind that if it doesn't appear on the leaderboard by the deadline this will resolve NO as well.

More related questions