Compare Claude Opus 4.5 by Anthropic against OpenAI Codex Mini by OpenAI, tested across 10 shared challenges. Updated February 2026.
10 challenges
Tests an AI's randomness and creativity
Here are 5 jokes:
Why don't scientists trust atoms? Because they make up everything.
What do you call a fish without eyes? A fsh.
Why did the scarecrow win an award? Because he was outstanding in his field.
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.
All 5 are unique - each joke has a different setup, punchline, and subject matter (atoms, fish, scarecrow, eyebrows, bear).
Sure—here are five jokes:
All five are distinct. Number of unique jokes: 5.
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
So I've been trying to drink more water. Everyone says drink more water. Eight glasses a day. And I'm doing it, I'm hydrated, I'm healthy, and I have to pee every forty-five minutes like I'm paying rent on the bathroom.
I was in a meeting last week—an important meeting—and I'm sitting there doing the math. "Okay, she's on slide 12, there's 40 slides, I can make it. I can make it." I could not make it. I did that thing where you stand up and just... leave. No explanation. Power move, honestly. Everyone probably thought I was so busy and important. Nope. Kidneys.
I turned 35 this year, and your body just starts... sending you notifications. Like a software update you keep ignoring. My knee cracked the other day and my wife looked at me like I'd been shot. I'm like, "No, that's just... that's just how I stand up now. From sitting. That's the sound of standing."
And doctors start asking you different questions at 35. "Any family history of—" YES. Whatever you're about to say, yes. Heart disease, diabetes, anxiety, bad decisions at buffets—it's all in there. My family tree is basically a WebMD search gone wrong.
I have a four-year-old, and nobody tells you that having kids means you're just gonna be wrong about everything for the rest of your life. My daughter corrected me on dinosaurs last week. I said "brontosaurus" and she hit me with, "Actually, it's apatosaurus, the brontosaurus was a misclassification."
She's FOUR. I don't even know where she learned that. I'm pretty sure I taught her the word "spoon" like eight months ago, and now she's out here peer-reviewing my paleontology.
Thank you, you've been a great crowd!
[Comedian walks onstage]
Hey, everyone! Great to see you all tonight. I'm Jamie—if you thought that intro was enthusiastic, you should've seen me when I realized this place has Wi-Fi.
Seriously, though, I love stand-up. For one thing, I get to share all the weird stuff I secretly do in my apartment… and instead of calling the cops, you guys just laugh. That's high-quality entertainment right there.
So, the other day I'm in the self-checkout lane at the grocery store—because I'm the kind of person who's thrilled by the thought of scanning my own kale. Of course, there's always that one person holding up everyone behind them because they can't figure out how to scan a cucumber. Meanwhile, the machine flashes: "UNEXPECTED ITEM IN BAGGING AREA." It's staring at me like it's personally offended that I tried to buy almonds in instead of potato chips. I'm over here trying to reason with it:
"Look, buddy, I'm just trying to eat slightly healthier. Isn't that what we all want?"
Then it asks me, "Do you want a produce discount?" The only discount I want is one that lets me leave this line five minutes ago. I press "No," and the machine glares at me with a red light and calls for help like I'm smuggling avocados under my shirt.
My phone has ruined me. I can't have any thought longer than three seconds without Instagram interrupting me. I'll be in the shower—fresh shampoo, head full of suds—and in my brain:
"Did my neighbor's cat post another selfie? I need to check."
Fifteen minutes later I'm drying off, scrolling through cats wearing sweaters. Why are cats always more photogenic than me? They don't even try. They're just like, "You feed me, you click 'Like.'" I wish I could have a boundary like that—"Hey, you gave me pizza once; you can go ahead and like my vacation photos."
Anyone here stuck in a group chat you can't escape? There's always that one friend who forwards the exact same video one hundred times, and everybody else thinks it's the highlight of their week. I just mute my phone like it's an ex sending memes at 2 a.m. Then I wake up at dawn, bleary-eyed, thinking, "What did I miss?" I open my phone—and it's Bob celebrating his new air-fryer chicken recipe for the fifth time. Meanwhile, I burned my eggs so badly this morning that my smoke alarm filed a restraining order against me.
I decided to get in shape. I downloaded one of those workout apps promising "six-pack abs in six weeks." After day one, I collected six new excuses why I can't work out:
Yesterday I finally ran so hard I passed out. When I came to, my neighbor was standing over me yelling, "You know you're not supposed to lean on lampposts, right?" I told him, "As soon as lampposts start offering aerodynamic support, I'll consider trusting them."
Alright, folks, that's my time. If you ever feel judged by a machine or betrayed by a group chat, come find me—I'll be the one in the self-checkout line, trying to convince a barcode scanner that kale is, in fact, a food. Thanks so much—you've been awesome!
Recreate an interactive, nostalgic Pokémon battle UI in a single HTML file.
Tests an AI's ability to create interactive web elements
Tests an AI's ability to create a detailed world map in SVG format
Generate a single-page, self-contained HTML webapp using Tailwind CSS for a randomly chosen category/industry/niche.
Generate SVG art of a randomly chosen animal in a setting of its choosing.
Recreate a classic, retro, macOS Desktop UI — make it interactive and nostalgic, make it as interactive as you can, you can be creative.. Feel free to showcase your skills. The entire application must be a single-page self-contained HTML file.
