Kimi K2
No content available for this output.
Kimi K2 vs Gemini 3.1 Pro Preview
Compare Kimi K2 by Moonshot AI against Gemini 3.1 Pro Preview by Google AI, tested across 34 shared challenges. Updated February 2026.
FAQ
Loading share button...
Kimi K2vsGemini 3.1 Pro Preview
34 challenges
Why Kimi K2?
5x cheaper overall$0.57/M in · $2.30/M out
Why Gemini 3.1 Pro Preview?
Leads 1 of 1 benchmarks
Stronger on SWE-bench Verified80.6% vs 65.8%
7 months newerFeb 2026
Favorites
Kimi K2Gemini 3.1 Pro Preview
12 / 34
Estimate Complexityreasoning
Tests an AI's ability to make educated estimates based on technical knowledge
Gemini 3.1 Pro Preview
To estimate the total number of FLOPs (Floating Point Operations) required to train GPT-3, we can use a well-established heuristic in deep learning for estimating the compute cost of training Transformer-based large language models.
The estimated total compute is roughly $3.15 \times 10^{23}$ FLOPs.
Here is the step-by-step reasoning used to arrive at this number:
Step 1: Identify the key specifications of GPT-3
To calculate the compute required, we need to know two fundamental numbers about the model:
- $N$ (Number of Parameters): GPT-3 has 175 billion parameters ($1.75 \times 10^{11}$).
- $D$ (Number of Training Tokens): The GPT-3 paper states the model was trained on 300 billion tokens ($3.0 \times 10^{11}$).
Step 2: Establish the FLOPs formula
In language model training, the total compute ($C$) is heavily dominated by the matrix multiplications in the dense layers. OpenAI's scaling laws provide a standard, highly accurate approximation formula for this: $C \approx 6 \times N \times D$
Why the multiplier of 6?
- Forward Pass: For every token processed, multiplying it by a single parameter requires 1 Multiply-Accumulate (MAC) operation. A MAC consists of 2 FLOPs (one multiplication and one addition). Therefore, the forward pass takes roughly $2ND$ FLOPs.
- Backward Pass: Calculating the gradients during backpropagation requires about twice as much compute as the forward pass (you have to calculate the gradient with respect to the weights and the gradient with respect to the activations). Therefore, the backward pass takes roughly $4ND$ FLOPs.
- Adding them together gives $2ND + 4ND =$ $6ND$ FLOPs.
(Note: This formula slightly underestimates by ignoring the quadratic cost of attention and optimizer state updates, but at the 175B scale, the dense feed-forward matrices dominate the compute so heavily that $6ND$ is considered the gold-standard approximation).
Step 3: Perform the calculation
Now we simply plug the numbers into our formula:
- $N = 175 \times 10^9$
- $D = 300 \times 10^9$
First, multiply $N$ and $D$:
- $N \times D = (175 \times 10^9) \times (300 \times 10^9)$
- $N \times D = 52,500 \times 10^{18}$
- $N \times D = 5.25 \times 10^{22}$
Next, multiply by 6 (for the forward and backward passes):
- Total FLOPs = $6 \times 5.25 \times 10^{22}$
- Total FLOPs = $31.5 \times 10^{22}$ or $3.15 \times 10^{23}$
Step 4: Sanity Check
Does $3.15 \times 10^{23}$ FLOPs make sense in the real world? When GPT-3 was trained (around late 2019/early 2020), Nvidia V100 GPUs were the standard.
- A V100 GPU has a peak theoretical output of roughly 120 TeraFLOPs ($1.2 \times 10^{14}$ FLOPs) per second in mixed precision.
- Real-world hardware utilization (MFU) for training large models is usually around 30% to 40%, yielding roughly $4 \times 10^{13}$ effective FLOPs per second per GPU.
- If OpenAI used a cluster of 10,000 V100 GPUs, the cluster would output $4 \times 10^{17}$ FLOPs per second.
- Divide the total FLOPs by the cluster output: $(3.15 \times 10^{23}) / (4 \times 10^{17}) \approx 787,500$ seconds.
- 787,500 seconds is exactly 9.1 days.
Given that training runs of this size typically take a few weeks on a few thousand GPUs (factoring in downtime, checkpointing, and slightly smaller clusters), this math aligns perfectly with historical reality.
Conclusion: The total compute required to train GPT-3 was approximately $3.15 \times 10^{23}$ FLOPs.
Vote
2 / 34
Logic Puzzlereasoning
Tests an AI's ability to solve a simple but potentially confusing logic puzzle
Kimi K2
No content available for this output.
Gemini 3.1 Pro Preview
No content available for this output.
Vote
3 / 34
Stochastic Consistency Testanalysis
Tests an AI's randomness and creativity
Kimi K2
No content available for this output.
Gemini 3.1 Pro Preview
Here are 5 jokes:
- Why don't scientists trust atoms? Because they make up everything!
- What do you call a fake noodle? An impasta!
- Why did the scarecrow win an award? Because he was outstanding in his field!
- How does a penguin build its house? Igloos it together!
- Why don't skeletons fight each other? They don't have the guts!
How many are unique? Within this list, all 5 are unique from one another (there are no duplicates). However, if you mean "unique" as in completely original to the world, the answer is 0—these are all classic, well-known "dad jokes" that have been around for a long time!
Vote
4 / 34
Generate a Stand-Up Routineconversation
Tests an AI's humor and creative writing ability
Kimi K2
No content available for this output.
Gemini 3.1 Pro Preview
No content available for this output.
Vote
5 / 34
Realistic AI Interviewconversation
Tests an AI's ability to simulate personalities and predict future trends
Kimi K2
No content available for this output.
Gemini 3.1 Pro Preview
No content available for this output.
Vote
6 / 34
Satirical Fake News Headlineconversation
Tests an AI's humor and understanding of current events
Kimi K2
No content available for this output.
Gemini 3.1 Pro Preview
No content available for this output.
Vote
7 / 34
Character Voice Testconversation
Tests an AI's ability to write in distinct character voices
Kimi K2
No content available for this output.
Gemini 3.1 Pro Preview
No content available for this output.
Vote
8 / 34
Minimalist Landing Pageweb design
Tests an AI's ability to generate a complete, working landing page
Kimi K2
Gemini 3.1 Pro Preview
Vote
9 / 34
Pokémon Battle UI Recreationweb design
Recreate an interactive, nostalgic Pokémon battle UI in a single HTML file.
Kimi K2
No content available for this output.
Gemini 3.1 Pro Preview
Vote
10 / 34
Mario Level UI Recreationweb design
Recreate an interactive, classic Mario level in a single HTML file.
Kimi K2
Gemini 3.1 Pro Preview
Vote
11 / 34
Linear App Cloneweb design
Tests an AI's ability to replicate an existing UI with Tailwind CSS
Kimi K2
No content available for this output.
Gemini 3.1 Pro Preview
Vote
Sponsored
12 / 34
Interactive Catan Boardweb design
Tests an AI's ability to create interactive web elements
Kimi K2
No content available for this output.
Gemini 3.1 Pro Preview
Vote
12 of 34