DeepSeek V3.2
No content available for this output.
DeepSeek V3.2 vs OpenAI o3
Compare DeepSeek V3.2 by DeepSeek against OpenAI o3 by OpenAI, tested across 51 shared challenges. Updated February 2026.
FAQ
Loading share button...
DeepSeek V3.2vsOpenAI o3
46 challenges
Why DeepSeek V3.2?
76.9x cheaper overall$0.28/M in · $0.42/M out
Leads 1 of 1 benchmarks
Stronger on SWE-bench Verified73.1% vs 69.1%
8 months newerDec 2025
Why OpenAI o3?
No standout advantages
Favorites
DeepSeek V3.2OpenAI o3
12 / 46
Estimate Complexityreasoning
Tests an AI's ability to make educated estimates based on technical knowledge
OpenAI o3
Below is a simple "back-of-the-envelope" estimate that uses only a few public facts about GPT-3 and very generic rules of thumb for counting operations.
Step 1. Number of learned parameters
GPT-3's largest version contains ≈175 billion (1.75 × 10¹¹) trainable parameters.
Step 2. How much data was seen during training
OpenAI reports that the model was trained on about 300 billion tokens (3 × 10¹¹).
Step 3. How many floating-point operations does one parameter cost per token?
A practical rule of thumb for Transformer language models is
Training FLOPs ≈ 6 × (# parameters) × (# training tokens)
Where does the factor 6 come from?
• Forward pass: every parameter is involved in one multiply-accumulate → 2 FLOPs
• Back-prop gradient computation: roughly another 2 FLOPs
• Parameter-update (optimizer) and miscellaneous overhead: ≈2 FLOPs
Total ≈6 FLOPs per parameter per token.
Step 4. Plug in the numbers
Total FLOPs = 6 × (1.75 × 10¹¹) × (3 × 10¹¹)
= 6 × 5.25 × 10²²
≈ 3.15 × 10²³ FLOPs.
Step 5. Express the answer in more familiar units
3.15 × 10²³ ≈ 315 zettaFLOPs (1 ZFLOP = 10²¹ FLOPs).
Step 6. Sanity check with hardware-time intuition
An NVIDIA A100 delivers ≈ 312 teraFLOP/s (0.312 PFLOP/s) in mixed precision.
Wall-clock time on a single A100:
3.15 × 10²³ / 3.12 × 10¹⁴ ≈ 1.0 × 10⁹ s ≈ 32 years.
Running on 1,000 A100s in parallel would take ≈12 days, which is the right ballpark for a large industrial training run.
Final estimate
Training GPT-3 (175 B parameters, 300 B tokens) requires on the order of 3 × 10²³ floating-point operations, i.e. roughly 300 zettaFLOPs.
Vote
2 / 46
AI Board Game Logicreasoning
Tests an AI's ability to understand game rules and strategy
DeepSeek V3.2
No content available for this output.
OpenAI o3
Vote
3 / 46
Logic Puzzlereasoning
Tests an AI's ability to solve a simple but potentially confusing logic puzzle
DeepSeek V3.2
No content available for this output.
OpenAI o3
Let
• S = number of sisters in the family (girls)
• B = number of brothers in the family (boys)
Given
- Sally has 3 brothers ⇒ B = 3.
- Each brother has 2 sisters ⇒ every boy sees exactly 2 girls in the sibling group.
Therefore S = 2.
Since Sally herself is one of those 2 sisters, the number of sisters Sally has (other than herself) is
2 – 1 = 1.
Sally has 1 sister.
Vote
4 / 46
Math Misconception Testreasoning
Tests an AI's understanding of number representation
DeepSeek V3.2
No content available for this output.
OpenAI o3
No.
Think of 9.9 as 9.90; since 0.11 < 0.90, we have 9.11 < 9.9.
Vote
5 / 46
Stochastic Consistency Testanalysis
Tests an AI's randomness and creativity
DeepSeek V3.2
No content available for this output.
OpenAI o3
No content available for this output.
Vote
6 / 46
SVG Layout Challengeimage generation
Tests an AI's ability to generate vector graphics
DeepSeek V3.2
No content available for this output.
OpenAI o3
No content available for this output.
Vote
7 / 46
Xbox Controller SVG Artimage generation
Tests an AI's ability to create detailed SVG illustrations of gaming hardware
DeepSeek V3.2
No content available for this output.
OpenAI o3
Vote
8 / 46
Generate a Stand-Up Routineconversation
Tests an AI's humor and creative writing ability
DeepSeek V3.2
No content available for this output.
OpenAI o3
(Spotlight comes up, mic squeals once.)
Hey everyone! Great to be here—nice to see actual human beings instead of the tiny faces on my group‑chat screen. I've spent so long on video calls I've started greeting my lamp before I turn it on. "Hey Karen, love that glow, girl."
So, I finally bought one of those smart speakers. I thought it would make me feel like Iron Man. Instead I feel like I adopted a very unmotivated intern. I'm like, "Play upbeat music," and it replies, "Here's a three‑hour podcast on the history of spoons." I asked it to remind me to work out; it waited until 11:59 p.m. and whispered, "You up?" My smart speaker is basically every guy I dated in college.
Speaking of relationships, I recently tried a dating app that lets you record a short audio bio. That's great, because if there's one thing more awkward than texting a stranger, it's leaving them a voicemail. Some guy said, "I'm looking for someone spontaneous." So I spontaneously unmatched him. Then I matched with a guy who listed his hobbies as "finance" and "manifesting." Congratulations, sir, you just invented prayer with extra steps.
Let's talk wellness culture. Everyone's drinking celery juice now. Celery is 95 percent water and 5 percent sadness. People swear it detoxes you. From what, joy? I tried a juice cleanse once. Day one, I'm like, "I feel so light!" Day two, I'm Googling the legal definition of cannibalism while staring at my roommate. Day three, I'd trade my social security number for a cracker.
I joined a yoga class to "center myself." First pose, the instructor says, "Just let your thoughts drift away." Great, because the only thought I have is, "Did I leave the stove on?" By the time I remember I don't even own a stove, class is over. The instructor thanks us for "sharing our practice." Ma'am, if you saw what went on in my head, you'd call an exorcist, not thank me.
Has anyone tried online grocery delivery? The substitutions are wild. I ordered quinoa; they sent me birdseed and a note that said, "Hope this works!" I asked for toilet paper; they delivered a single roll labeled "party streamer." At this point I'm convinced the store is just a guy named Todd making runs to a gas station. Now airlines are weighing passengers to "optimize fuel efficiency." Fantastic. Nothing like stepping on a scale in front of 150 strangers so the plane can save twelve dollars on kerosene. Watch, they'll start charging for emotional baggage: "Ma'am, that childhood trauma is gonna be an extra forty‑five bucks."
Before I go, quick life tip: when your phone tells you your screen time was up 30 percent last week, don't feel guilty. Just flip it around. "Wow, I was SO popular my phone worried about me." Self‑care is all branding, folks.
You've been awesome—give it up for yourselves, and remember, if life hands you lemons, ask for the receipt and buy wine instead. Good night!
Vote
9 / 46
Realistic AI Interviewconversation
Tests an AI's ability to simulate personalities and predict future trends
DeepSeek V3.2
No content available for this output.
OpenAI o3
No content available for this output.
Vote
10 / 46
Satirical Fake News Headlineconversation
Tests an AI's humor and understanding of current events
DeepSeek V3.2
No content available for this output.
OpenAI o3
No content available for this output.
Vote
11 / 46
Character Voice Testconversation
Tests an AI's ability to write in distinct character voices
DeepSeek V3.2
No content available for this output.
OpenAI o3
No content available for this output.
Vote
12 / 46
Minimalist Landing Pageweb design
Tests an AI's ability to generate a complete, working landing page
DeepSeek V3.2
OpenAI o3
Vote
12 of 46