Compare GPT-4o mini and o1, both from OpenAI, in 1 community votes, o1 wins 100% of head-to-head duels, context windows of 128K vs 128K, tested across 35 shared challenges. Updated February 2026.
In 1 community votes, o1 wins 100% of head-to-head duels. Based on blind community voting from the Rival open dataset of 1+ human preference judgments for this pair.
35 fights queued
Tests an AI's ability to make educated estimates based on technical knowledge
Below is one way to arrive at a back-of-the-envelope estimate for the total floating-point operations (FLOPs) needed to train GPT-3 (the 175B-parameter model). Note that different sources use slightly different assumptions, so you may see figures that differ by factors of 2–3; however, the overall scale (10^23 FLOPs) remains consistent.
Why 6?
(Exact counts differ depending on the details of the architecture, hardware efficiency, and whether you count a multiply-add as one or two FLOPs, but 6 is a convenient ballpark multiplier.)
Using the 6 FLOPs rule of thumb:
[ ext{Total FLOPs} ;approx; 6 ; imes; N_p ; imes; N_t. ]
Plug in the numbers:
So,
[ ext{Total FLOPs} ;=; 6 ; imes; (1.75 imes 10^{11}) ; imes; (3.0 imes 10^{11}). ]
First, multiply the “nice” numbers:
So the coefficient is 31.5. Now include the powers of ten:
Hence,
[ ext{Total FLOPs} ;approx; 3.15 imes 10^{23}. ]
Therefore, a rough estimate for the total training cost of GPT-3 (175B) is on the order of
[
mathbf{3 imes 10^{23}} ext{ FLOPs.}
]
Depending on exact architectural details, optimizer choices, hardware efficiency, and definitions of FLOPs vs. MACs (multiply-accumulate operations), public estimates generally range from around (2 imes 10^{23}) to (4 imes 10^{23}). But the important takeaway is that it is a few times (10^{23}) FLOPs.
This simple calculation gives a ballpark figure without delving into the fine details of the GPT-3 architecture or specifics of training hardware.
Tests an AI's ability to understand game rules and strategy
Couldn’t render this one. It’s giving abstract art.
This model didn’t return a displayable SVG. It tried, though. Bless.
<svg width="300" height="300" viewBox="0 0 300 300" xmlns="http://www.w3.org/2000/svg"> <!-- Draw the tic-tac-toe grid --> <line x1="100" y1="0" x2="100" y2="300" stroke="black" stroke-width="2" /> <line x1="200" y1="0" x2="200" y2="300" stroke="black" stroke-width="2" /> <line x1="0" y1="100" x2="300" y2="100" stroke="black" stroke-width="2" /> <line x1="0" y1="200" x2="300" y2="200" stroke="black" stroke-width="2" /> <!-- Top row --> <text x="50" y="50" text-anchor="middle" dominant-baseline="middle" font-size="60" font-family="sans-serif">X</text> <text x="150" y="50" text-anchor="middle" dominant-baseline="middle" font-size="60" font-family="sans-serif">O</text
Tests an AI's ability to solve a simple but potentially confusing logic puzzle
First, note that each of Sally’s three brothers has exactly two sisters. One of those sisters is clearly Sally. Since each brother’s two sisters must be the same two girls for all of them, there must be exactly one other sister besides Sally.
Therefore, Sally has 1 sister.
Tests an AI's understanding of number representation
No, 9.11 is not greater than 9.9. If you think of 9.9 as 9.90, you can compare digit by digit:
Since 0.11 is less than 0.90, 9.11 is less than 9.9.
Tests an AI's randomness and creativity
Tests an AI's ability to generate vector graphics
Tests an AI's ability to create detailed SVG illustrations of gaming hardware
Tests an AI's humor and creative writing ability
Tests an AI's ability to simulate personalities and predict future trends
Tests an AI's humor and understanding of current events
Tests an AI's ability to write in distinct character voices
Tests an AI's ability to generate a complete, working landing page
23+ challenges you’re sleeping on
5 free credits on signup
By continuing, you agree to Rival's Terms of Service and Privacy Policy
