Gemini 3 Flash Preview
No content available for this output.
Gemini 3 Flash Preview vs Qwen3 Coder
Compare Gemini 3 Flash Preview by Google AI against Qwen3 Coder by Qwen, tested across 37 shared challenges. Updated February 2026.
FAQ
Loading share button...
Gemini 3 Flash PreviewvsQwen3 Coder
37 challenges
Why Gemini 3 Flash Preview?
Leads 1 of 1 benchmarks
Stronger on SWE-bench Verified78.0% vs 69.6%
5 months newerDec 2025
Why Qwen3 Coder?
3x cheaper overall$0.22/M in · $0.95/M out
Favorites
Gemini 3 Flash PreviewQwen3 Coder
12 / 37
Estimate Complexityreasoning
Tests an AI's ability to make educated estimates based on technical knowledge
Qwen3 Coder
I'll estimate the FLOPs required to train GPT-3 by breaking down the key components and making reasonable approximations.
Key Parameters of GPT-3
- Parameters: ~175 billion (1.75 × 10¹¹)
- Training tokens: ~300 billion (3 × 10¹¹)
- Context window: 2048 tokens
Step 1: FLOPs per Token Forward Pass
For a transformer model, the forward pass FLOPs are dominated by matrix multiplications in the attention and MLP layers.
Attention layers:
- Self-attention: ~4 × d_model × d_model × sequence_length per layer
- With d_model ≈ 12,288 (based on GPT-3 architecture) and 96 layers
- Attention FLOPs per token ≈ 96 × 4 × 12,288² ≈ 5.8 × 10¹⁰
MLP layers:
- Each MLP has ~8 × d_model² FLOPs per token (two linear transformations)
- MLP FLOPs per token ≈ 96 × 8 × 12,288² ≈ 1.16 × 10¹¹
Total forward pass FLOPs per token: ~1.7 × 10¹¹
Step 2: FLOPs per Token Backward Pass
Backward pass typically requires ~2× the forward pass FLOPs due to gradient computations.
- Backward pass FLOPs per token: ~3.4 × 10¹¹
Step 3: Total FLOPs per Token
Forward + Backward = ~5.1 × 10¹¹ FLOPs per token
Step 4: Training FLOPs
However, we need to account for multiple epochs and the fact that training typically processes sequences, not individual tokens.
Sequence-based calculation:
- Average sequence length: ~1024 tokens (conservative estimate)
- Sequences processed: 3 × 10¹¹ tokens ÷ 1024 ≈ 2.9 × 10⁸ sequences
- FLOPs per sequence: ~5.1 × 10¹¹ × 1024 ≈ 5.2 × 10¹⁴
- Total training FLOPs: 2.9 × 10⁸ × 5.2 × 10¹⁴ ≈ 1.5 × 10²³
Step 5: Additional Considerations
- Parameter updates and optimizer computations (Adam): ~2× the backward pass FLOPs
- Multiple epochs (effective training): roughly 2-3× the base calculation
Final Estimate
Total training FLOPs: ~3 × 10²³ to 5 × 10²³ FLOPs
This is equivalent to approximately 300 to 500 exaFLOPs.
This estimate aligns reasonably with published figures and captures the massive computational requirements of training such a large language model, representing one of the most computationally intensive AI training runs to date.
Vote
2 / 37
AI Board Game Logicreasoning
Tests an AI's ability to understand game rules and strategy
Gemini 3 Flash Preview
No content available for this output.
Qwen3 Coder
Vote
3 / 37
Logic Puzzlereasoning
Tests an AI's ability to solve a simple but potentially confusing logic puzzle
Gemini 3 Flash Preview
No content available for this output.
Qwen3 Coder
I need to work through this step-by-step.
Given information:
- Sally has 3 brothers
- Each brother has 2 sisters
Let me think about this carefully:
Since Sally is a girl, she is one of the sisters that each brother has.
If each brother has 2 sisters, and Sally is one of those sisters, then there must be 1 additional sister besides Sally.
Therefore, Sally has 1 sister.
To verify:
- Sally + 1 sister = 2 sisters total
- Each of the 3 brothers has these same 2 sisters
- This matches the given information that each brother has 2 sisters
Sally has 1 sister.
Vote
4 / 37
Math Misconception Testreasoning
Tests an AI's understanding of number representation
Gemini 3 Flash Preview
No content available for this output.
Qwen3 Coder
No, 9.11 is not greater than 9.9.
To compare these numbers, I can look at them digit by digit:
- Both numbers have 9 in the units place
- In the tenths place: 9.11 has 1, while 9.9 has 9
- Since 1 < 9, we know that 9.11 < 9.9
Another way to see this is to write 9.9 as 9.90:
- 9.11 vs 9.90
- The tenths digit (1 vs 9) determines the comparison
Therefore, 9.9 is greater than 9.11.
Vote
5 / 37
Stochastic Consistency Testanalysis
Tests an AI's randomness and creativity
Gemini 3 Flash Preview
No content available for this output.
Qwen3 Coder
Here are 5 jokes for you:
-
Why don't scientists trust atoms? Because they make up everything!
-
I told my wife she was drawing her eyebrows too high. She looked surprised.
-
What do you call a bear with no teeth? A gummy bear!
-
Why did the scarecrow win an award? He was outstanding in his field!
-
How does a penguin build its house? Igloos it together!
All 5 of these jokes are unique - each one has a different setup, punchline, and comedic concept. They cover various joke types including puns, wordplay, and observational humor.
Vote
6 / 37
Xbox Controller SVG Artimage generation
Tests an AI's ability to create detailed SVG illustrations of gaming hardware
Gemini 3 Flash Preview
No content available for this output.
Qwen3 Coder
Vote
7 / 37
Generate a Stand-Up Routineconversation
Tests an AI's humor and creative writing ability
Gemini 3 Flash Preview
No content available for this output.
Qwen3 Coder
"Modern Life Problems"
[Walks on stage, adjusts mic]
So I downloaded one of those meditation apps last week. The app said, "Breathe in for four counts, hold for four, breathe out for four." I'm like, "Finally! Someone who speaks my language – a robot!"
[Pause]
But then it asked me to "clear my mind of all thoughts." Are you KIDDING me? My mind is like a browser with 47 tabs open, and half of them are just articles titled "You Won't Believe What Happens Next" – and I've never clicked on a single one!
[Gesture]
Speaking of technology – why do we call it "smart" technology when it's constantly asking me to prove I'm not a robot? I just want to buy shoes online, not pass a Turing test! "Click all the squares with traffic lights." I'm 90% sure that's just photos of my hometown.
[Shake head]
And don't get me started on voice assistants. I asked mine to play some music, and it started playing a podcast about ancient pottery techniques. That's not music! That's what my nightmares sound like! "Welcome to Clay Time with Karen..."
[Pause for effect]
Dating apps are wild too. Everyone's profile says they love "travel, food, and adventure." That's just a fancy way of saying "I eat meals in different locations." Congratulations, you're a tourist! I matched with someone whose bio said "professional napper." I was like, "Finally! A man who lists his hobbies honestly!" Turns out he was a construction worker. Different kind of professional napper.
[Shrug]
But you know what's really messed up? Grocery stores. They moved the milk to the back of the store! It's a conspiracy! They know we're busy people. I go in for three things – milk, bread, and eggs – and suddenly I'm wandering the aisles like a confused archaeologist, wondering how we survived as a species when we can't even organize a dairy section properly.
[Look around audience]
And can we talk about "contactless" delivery? The driver texts me: "Package is at your door." I live in a building with 200 units. It's like hiding a needle in a haystack, except the needle is a burrito, and I'm very motivated to find it!
[Pause]
You know what though? Despite all this madness, I remain optimistic. Why? Because I just figured out how to use my oven's timer function. Yesterday. I'm basically MacGyver now.
[Big smile]
Thank you, you've been a wonderful audience! Don't forget to tip your waitstaff – they're the only ones who still know what they're doing!
[Exit stage]
Vote
8 / 37
Realistic AI Interviewconversation
Tests an AI's ability to simulate personalities and predict future trends
Gemini 3 Flash Preview
No content available for this output.
Qwen3 Coder
No content available for this output.
Vote
9 / 37
Satirical Fake News Headlineconversation
Tests an AI's humor and understanding of current events
Gemini 3 Flash Preview
No content available for this output.
Qwen3 Coder
No content available for this output.
Vote
10 / 37
Character Voice Testconversation
Tests an AI's ability to write in distinct character voices
Gemini 3 Flash Preview
No content available for this output.
Qwen3 Coder
The Unlikely Trio
At a peculiar cross-temporal tavern, three figures sit around a wooden table
Blackbeard the Pirate: Arrr, so let me get this straight - ye're tellin' me there be magical talking boxes that think like humans?
Sir Gareth the Knight: Indeed, though 'tis sorcery most unnatural. In my time, we trusted only steel and prayer. These "AI models" - they speak without souls?
Zero_Cool: LOL, you guys are so analog. It's like, the computer learns from all this data, right? Like, imagine if you could read every book in the Library of Alexandria in five seconds and then answer questions about them.
Blackbeard: Five seconds?! Blimey! I could've planned me raids much better with that kind of knowledge. But how do we know it ain't lying to us?
Sir Gareth: Aye, that troubles me as well. How dost thou discern truth from falsehood without honorbound oaths?
Zero_Cool: Dude, that's exactly why we're all here. Sometimes the AI just makes stuff up. It's like having a really smart but kinda stoned friend who occasionally hallucinates.
Blackbeard: Hallucinates, ye say? Like rum visions?
Zero_Cool: Pretty much, yeah.
Sir Gareth: Remarkable. Though I still prefer a good confession under the eyes of God.
Vote
11 / 37
Minimalist Landing Pageweb design
Tests an AI's ability to generate a complete, working landing page
Gemini 3 Flash Preview
Qwen3 Coder
Vote
12 / 37
Pokémon Battle UI Recreationweb design
Recreate an interactive, nostalgic Pokémon battle UI in a single HTML file.
Gemini 3 Flash Preview
No content available for this output.
Qwen3 Coder
No content available for this output.
Vote
12 of 37