Will Gemini achieve a score above 90% on the MATH benchmark?
Standard
20
แน€4926
resolved Sep 16
Resolved
YES

The current SOTA is 84.3% from GPT-4 Code Interpreter. Code & tool use is allowed.

Get
แน€1,000
and
S1.00
Sort by:

I should have specified the exact model. What I intended was the first Gemini 1.0 family, not the entire Gemini series. My bad guys. Since the question itself can be interpreted as the Gemini series, so I just resolve this to Yes.

Since this market has no restrictions on public availability or zero shot, I think this should probably already resolve as yes per Gemini 1.5 report

This is a separate MATH than the one that Google reported the benchmark on. And I don't see it beating GPT4 by so much, given most of the other scores were very close.

limit order for yes 10