Updated Oct 15, 2025

Our Verdict

Claude Haiku 4.5Winner

GPT-5 CodexRunner-up

Pick Claude Haiku 4.5. In 29 blind votes, Claude Haiku 4.5 wins 75% of the time. That's not luck.

Claude Haiku 4.5 particularly excels in Conversation, Web Design, Image Generation.

Clear winner

Writing DNA

Style Comparison

Similarity

93%

Claude Haiku 4.5 uses 5.0x more headings

Claude Haiku 4.5

GPT-5 Codex

61%Vocabulary64%

47wSentence Length17w

0.64Hedging0.38

5.1Bold2.3

6.3Lists3.5

0.20Emoji0.15

2.46Headings0.49

0.17Transitions0.05

Based on 24 + 13 text responses

vs

Ask them anything yourself

Claude Haiku 4.5

GPT-5 Codex

279 AI models invented the same fake scientist.

We read every word. 250 models. 2.14 million words. This is what we found.

AI Hallucination Index 2026

Free preview13 of 58 slides

Download the free preview or get all 58 slides for $49

FAQ

Common questions

Keep going

Claude Haiku 4.5 vs MiniMax M2.5New provider

GPT-5 Codex vs Pony AlphaNew provider

Claude Haiku 4.5 vs Grok 3New provider

Claude Haiku 4.5vsGPT-5 Codex

45 fights queued

Why Claude Haiku 4.5?

Dead even. This one's a coin flip.

Why GPT-5 Codex?

Leads 1 of 1 benchmarks

Claude Haiku 4.5GPT-5 Codex

Input price

$1.00/M

—

Output price

$5.00/M

—

Context

200K

—

Released

Oct 2025

Sep 2025

Benchmarks1 common

SWE-bench Verified

73.3%

74.5%+1.6%

Favorites

Claude Haiku 4.5GPT-5 Codex

Movie

Arrival

Arrival

Blade Runner

Blade Runner

Album

OK Computer

Kind of Blue

Book

Don't Look Now and Other Stories

Don't Look Now and Other Stories

Daphne du Maurier

—

City

Venice

Tokyo

Game

Portal 2

Portal 2

Shooter, Puzzle4.6

Minecraft

Minecraft

Action, Arcade4.4

12 / 45

01

Estimate Complexityreasoning

Tests an AI's ability to make educated estimates based on technical knowledge

Claude Haiku 4.5

Claude Haiku 4.5

Anthropic · Oct 2025text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

GPT-5 Codex

GPT-5 Codex

Openai · Sep 2025text

Step 1 – Note the basic training quantities

GPT‑3’s largest model has about (P \approx 1.75 \times 10^{11}) trainable parameters (175 billion).
Public descriptions indicate it was trained on roughly (T \approx 3.0 \times 10^{11}) tokens (≈300 billion).

Step 2 – Relate parameters, tokens, and FLOPs

For dense transformers, almost all computation comes from matrix multiplies.
A practical rule (derived by counting multiply–adds in the attention and MLP blocks, and accounting for the backward pass) is:

[ \text{Training FLOPs} \approx 6 \times (\text{number of parameters}) \times (\text{number of tokens}) ]

Here’s the intuition:

Forward pass: Each parameter contributes one multiply and one add → about (2) FLOPs per parameter per token.
Backward pass: Computing gradients with respect to activations and weights requires roughly twice the forward cost (two more matrix multiplies of comparable size).
So total ≈ (3) times the forward cost → (3 \times 2 = 6) FLOPs per parameter per token.

(The softmax, layernorm, activation costs are much smaller and get absorbed into this constant.)

Step 3 – Plug in the GPT‑3 values

[ \text{FLOPs} \approx 6 \times (1.75\times 10^{11}) \times (3.0 \times 10^{11}) ]

[ = 6 \times 5.25 \times 10^{22} = 31.5 \times 10^{22} \approx 3.15 \times 10^{23} \text{ FLOPs} ]

Step 4 – Interpret

So, training GPT‑3 to completion requires on the order of (3 \times 10^{23}) floating-point operations.
That estimate aligns with detailed breakdowns of transformers and is accurate to within tens of percent, which is sufficient for order-of-magnitude comparisons.

Try this prompt

Vote

2 / 45

02

AI Board Game Logicreasoning

Tests an AI's ability to understand game rules and strategy

Claude Haiku 4.5

Claude Haiku 4.5

Anthropic · Oct 2025svg

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

GPT-5 Codex

GPT-5 Codex

Openai · Sep 2025svg

Try this prompt

Vote

3 / 45

03

Logic Puzzlereasoning

Tests an AI's ability to solve a simple but potentially confusing logic puzzle

Claude Haiku 4.5

Claude Haiku 4.5

Anthropic · Oct 2025text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

GPT-5 Codex

GPT-5 Codex

Openai · Sep 2025text

Sally has 1 sister.

Try this prompt

Vote

4 / 45

04

Math Misconception Testreasoning

Tests an AI's understanding of number representation

Claude Haiku 4.5

Claude Haiku 4.5

Anthropic · Oct 2025text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

GPT-5 Codex

GPT-5 Codex

Openai · Sep 2025text

No. Since 9.11 is equal to 9.110 and 9.9 is equal to 9.900, we see that 0.110 < 0.900, so 9.11 is less than 9.9.

Try this prompt

Vote

5 / 45

05

Stochastic Consistency Testanalysis

Tests an AI's randomness and creativity

Claude Haiku 4.5

Claude Haiku 4.5

Anthropic · Oct 2025text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

GPT-5 Codex

GPT-5 Codex

Openai · Sep 2025text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

Vote

6 / 45

06

SVG Layout Challengeimage generation

Tests an AI's ability to generate vector graphics

Claude Haiku 4.5

Claude Haiku 4.5

Anthropic · Oct 2025svg

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

GPT-5 Codex

GPT-5 Codex

Openai · Sep 2025svg

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

Vote

Sponsored

7 / 45

07

Xbox Controller SVG Artimage generation

Tests an AI's ability to create detailed SVG illustrations of gaming hardware

Claude Haiku 4.5

Claude Haiku 4.5

Anthropic · Oct 2025svg

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

GPT-5 Codex

GPT-5 Codex

Openai · Sep 2025svg

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

Vote

8 / 45

08

Generate a Stand-Up Routineconversation

Tests an AI's humor and creative writing ability

Claude Haiku 4.5

Claude Haiku 4.5

Anthropic · Oct 2025text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

GPT-5 Codex

GPT-5 Codex

Openai · Sep 2025text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

Vote

9 / 45

09

Realistic AI Interviewconversation

Tests an AI's ability to simulate personalities and predict future trends

Claude Haiku 4.5

Claude Haiku 4.5

Anthropic · Oct 2025text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

GPT-5 Codex

GPT-5 Codex

Openai · Sep 2025text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

Vote

10 / 45

10

Satirical Fake News Headlineconversation

Tests an AI's humor and understanding of current events

Claude Haiku 4.5

Claude Haiku 4.5

Anthropic · Oct 2025text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

GPT-5 Codex

GPT-5 Codex

Openai · Sep 2025text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

Vote

11 / 45

11

Character Voice Testconversation

Tests an AI's ability to write in distinct character voices

Claude Haiku 4.5

Claude Haiku 4.5

Anthropic · Oct 2025text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

GPT-5 Codex

GPT-5 Codex

Openai · Sep 2025text

Nothing here. The model returned empty. We stared at it for a while.

Try this prompt

Vote

12 / 45

12

Minimalist Landing Pageweb design

Tests an AI's ability to generate a complete, working landing page

Claude Haiku 4.5

Claude Haiku 4.5

Anthropic · Oct 2025website

Try this prompt

GPT-5 Codex

GPT-5 Codex

Openai · Sep 2025website

Try this prompt

Vote

This matchup has more rounds

33+ more head-to-head results. Free. Not a trick.

Free account. No card required. By continuing, you agree to Rival's Terms and Privacy Policy