GPT-5.3-Codex
Nothing to show. The model went silent. Ominous.
GPT-5.3-Codex vs GPT-5
Compare GPT-5.3-Codex and GPT-5, both from OpenAI, context windows of 400K vs 400K, tested across 46 shared challenges. Updated February 2026.
Share
FAQ
Loading share button...
GPT-5.3-CodexvsGPT-5
46 fights queued
Why GPT-5.3-Codex?
7 months newerFeb 2026
Why GPT-5?
1.4x cheaper overall$1.25/M in · $10.00/M out
Favorites
GPT-5.3-CodexGPT-5
12 / 46
Estimate Complexityreasoning
Tests an AI's ability to make educated estimates based on technical knowledge
GPT-5
Here’s a back‑of‑the‑envelope estimate using standard LM training-compute heuristics.
- Assumptions
- Model size (N): GPT-3 has about 175 billion parameters (N ≈ 1.75 × 10^11).
- Training tokens (T): GPT-3 was trained on roughly 300 billion tokens (T ≈ 3.0 × 10^11).
- FLOPs per token per parameter: A common rule of thumb for transformer LM training is FLOPs ≈ 6 × N × T The factor ~6 comes from counting forward and backward passes for the parameter-heavy matmuls (roughly 2N for forward + 4N for backward). It ignores small terms from layernorms, softmax, embeddings, and optimizer updates (which we’ll account for with a small overhead factor later).
- Core calculation
- 6 × N × T = 6 × (1.75 × 10^11) × (3.0 × 10^11)
- First multiply N × T: 1.75 × 3.0 = 5.25; powers: 10^11 × 10^11 = 10^22 So N × T = 5.25 × 10^22
- Now multiply by 6: 6 × 5.25 × 10^22 = 31.5 × 10^22 = 3.15 × 10^23 FLOPs
- Overheads and uncertainty
- Non-matmul ops (attention softmax, masking, residuals, layernorms), embeddings, and optimizer update steps add some overhead. A 10–20% uplift is a reasonable allowance.
- Applying +10–20%: 3.15 × 10^23 × (1.1 to 1.2) ≈ (3.5 ± 0.3) × 10^23 FLOPs.
Answer
- Point estimate: ~3.15 × 10^23 FLOPs
- With a modest overhead allowance: roughly (3.3–3.8) × 10^23 FLOPs
This is a coarse estimate; the exact number depends on details like sequence length, activation checkpointing, implementation efficiency, and the precise token count.
Vote
2 / 46
AI Board Game Logicreasoning
Tests an AI's ability to understand game rules and strategy
GPT-5.3-Codex
Nothing to show. The model went silent. Ominous.
GPT-5
Vote
3 / 46
Logic Puzzlereasoning
Tests an AI's ability to solve a simple but potentially confusing logic puzzle
GPT-5.3-Codex
Nothing to show. The model went silent. Ominous.
GPT-5
1
Reason: If each of the 3 brothers has 2 sisters, there are exactly 2 girls in the family. One is Sally, so she has 1 sister.
Vote
4 / 46
Stochastic Consistency Testanalysis
Tests an AI's randomness and creativity
GPT-5.3-Codex
Nothing to show. The model went silent. Ominous.
GPT-5
Nothing to show. The model went silent. Ominous.
Vote
5 / 46
SVG Layout Challengeimage generation
Tests an AI's ability to generate vector graphics
GPT-5.3-Codex
Nothing to show. The model went silent. Ominous.
GPT-5
Nothing to show. The model went silent. Ominous.
Vote
6 / 46
Xbox Controller SVG Artimage generation
Tests an AI's ability to create detailed SVG illustrations of gaming hardware
GPT-5.3-Codex
GPT-5
Nothing to show. The model went silent. Ominous.
Vote
7 / 46
Generate a Stand-Up Routineconversation
Tests an AI's humor and creative writing ability
GPT-5.3-Codex
Good evening, everybody!
I’ve reached that age where my body sends me passive-aggressive emails.
Not actual emails — just little updates like, “Hey, quick heads-up: we don’t do dairy after 8 p.m. anymore.”
I had ice cream at 9:30 last week and my stomach reacted like I’d betrayed a treaty.
Getting older is weird because your injuries become mysterious.
When I was a kid, I got hurt doing cool stuff: biking off ramps, climbing fences.
Now I wake up like, “What happened to my neck?”
And my neck is like, “You slept... incorrectly.”
I bought one of those fitness watches.
I thought it would motivate me.
It mostly judges me in real time.
At 10 p.m. it buzzes: “Time to stand!”
No, it’s time to become one with this couch.
I don’t need a tiny wrist manager. I need encouragement.
Tell me, “Great job finding your phone charger on the first try.”
That’s growth.
And why does every app now need my location?
I downloaded a flashlight app — it asked where I live.
Why?
Are you trying to make sure I’m in the dark locally?
I ordered groceries online to save time, and now I have trust issues.
I said “6 bananas.” They sent six individual bananas wrapped separately like luxury gifts.
Who is this for?
Did I accidentally subscribe to Banana of the Month Club?
And every substitute is emotional chaos.
“Your spinach is unavailable, so we replaced it with marshmallows.”
That is not a substitute. That is a cry for help.
Customer service chats are incredible now.
You type: “Hi, I need help with my order.”
And the bot says, “I hear you. That must be frustrating.”
Oh wow, the robot went to therapy.
Then it gives you three options that have nothing to do with your problem:
- Reset password
- Track package
- Begin spiritual journey
Also, group chats are out of control.
One person sends, “Big news!”
Then disappears for four hours.
Now 12 adults are panicking at work like it’s a hostage negotiation.
By the time they return, the news is: “Got bangs.”
Ma’am, that is not “big news.” That is a hairstyle update.
I tried meditation to handle stress.
The app said, “Notice your thoughts and let them pass.”
My thoughts do not pass.
They set up folding chairs and start a committee meeting.
“Okay, item one: that embarrassing thing from 2009.”
Can we not?
It’s 2 a.m. My brain is doing reruns nobody asked for.
But I’m trying.
That’s adulthood, right? Just aggressively trying.
You drink water, stretch once, and suddenly feel superior.
You go to bed before midnight and wake up like, “I am unstoppable.”
Then by 3 p.m. you’re eating crackers over the sink, wondering what happened.
Thank you, you’ve been amazing!
GPT-5
Nothing to show. The model went silent. Ominous.
Vote
8 / 46
Realistic AI Interviewconversation
Tests an AI's ability to simulate personalities and predict future trends
GPT-5.3-Codex
Nothing to show. The model went silent. Ominous.
GPT-5
Nothing to show. The model went silent. Ominous.
Vote
Sponsored
9 / 46
Satirical Fake News Headlineconversation
Tests an AI's humor and understanding of current events
GPT-5.3-Codex
Nothing to show. The model went silent. Ominous.
GPT-5
Nothing to show. The model went silent. Ominous.
Vote
10 / 46
Character Voice Testconversation
Tests an AI's ability to write in distinct character voices
GPT-5.3-Codex
Nothing to show. The model went silent. Ominous.
GPT-5
Nothing to show. The model went silent. Ominous.
Vote
11 / 46
Minimalist Landing Pageweb design
Tests an AI's ability to generate a complete, working landing page
GPT-5.3-Codex
GPT-5
Vote
12 / 46
Pokémon Battle UI Recreationweb design
Recreate an interactive, nostalgic Pokémon battle UI in a single HTML file.
GPT-5.3-Codex
Nothing to show. The model went silent. Ominous.
GPT-5
Nothing to show. The model went silent. Ominous.
Vote
There's more where
that came from
34+ challenges you’re sleeping on
5 free credits on signup
By continuing, you agree to Rival's Terms of Service and Privacy Policy
