Updated Apr 2, 2026

Our Verdict

Google: Gemma 4 31B

GPT-5

No community votes yet. On paper, these are closely matched - try both with your actual task to see which fits your workflow.

Google: Gemma 4 31B is 25x cheaper per token — worth considering if cost matters.

Too close to call

Writing DNA

Style Comparison

Similarity

95%

Google: Gemma 4 31B uses 23.8x more emoji

Google: Gemma 4 31B

GPT-5

57%Vocabulary65%

18wSentence Length13w

0.37Hedging0.20

7.1Bold0.2

4.0Lists4.6

0.24Emoji0.00

0.94Headings0.16

0.16Transitions0.03

Based on 23 + 16 text responses

vs

Ask them anything yourself

Google: Gemma 4 31B

GPT-5

FAQ

Keep going

Google: Gemma 4 31B vs Llama 4 MaverickNew provider

Google: Gemma 4 31B vs Claude Opus 4New provider

Google: Gemma 4 31B vs MiniMax M1New provider

Google: Gemma 4 31BvsGPT-5

38 fights queued

Why Google: Gemma 4 31B?

22x cheaper overall$0.14/M in · $0.40/M out

8 months newerApr 2026

Why GPT-5?

1.5x more context400K

Google: Gemma 4 31BGPT-5

Input price

$0.14/M

$1.25/M

Output price

$0.40/M

$10.00/M

Context

262K

400K

Released

Apr 2026

Aug 2025

Favorites

Google: Gemma 4 31BGPT-5

Movie

Her

Her

—

Album

Kind of Blue

—

Book

Gödel, Escher, Bach

Gödel, Escher, Bach

Douglas R. Hofstadter

—

City

Tokyo

—

Game

The Witness

The Witness

Indie, Adventure4.0

—

12 / 38

01

Estimate Complexityreasoning

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

Google: Gemma 4 31B

Google: Gemma 4 31B

Google · Apr 2026text

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

Try this prompt

GPT-5

GPT-5

Openai · Aug 2025text

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.

Try this prompt

Vote

2 / 38

02

AI Board Game Logicreasoning

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

Google: Gemma 4 31B

Google: Gemma 4 31B

Google · Apr 2026svg

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

Try this prompt

GPT-5

GPT-5

Openai · Aug 2025svg

Try this prompt

Vote

3 / 38

03

Logic Puzzlereasoning

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

Google: Gemma 4 31B

Google: Gemma 4 31B

Google · Apr 2026text

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

Try this prompt

GPT-5

GPT-5

Openai · Aug 2025text

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.

Try this prompt

Vote

4 / 38

04

Xbox Controller SVG Artimage generation

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

Google: Gemma 4 31B

Google: Gemma 4 31B

Google · Apr 2026svg

Try this prompt

GPT-5

GPT-5

Openai · Aug 2025svg

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

Try this prompt

Vote

5 / 38

05

Realistic AI Interviewconversation

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

Google: Gemma 4 31B

Google: Gemma 4 31B

Google · Apr 2026text

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

Try this prompt

GPT-5

GPT-5

Openai · Aug 2025text

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

Try this prompt

Vote

6 / 38

06

Satirical Fake News Headlineconversation

Tests an AI's humor and understanding of current events

Google: Gemma 4 31B

Google: Gemma 4 31B

Google · Apr 2026text

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

Try this prompt

GPT-5

GPT-5

Openai · Aug 2025text

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

Try this prompt

Vote

Sponsored

7 / 38

07

Character Voice Testconversation

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

Google: Gemma 4 31B

Google: Gemma 4 31B

Google · Apr 2026text

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

Try this prompt

GPT-5

GPT-5

Openai · Aug 2025text

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

Try this prompt

Vote

8 / 38

08

Pokémon Battle UI Recreationweb design

Recreate an interactive, nostalgic Pokémon battle UI in a single HTML file.

Google: Gemma 4 31B

Google: Gemma 4 31B

Google · Apr 2026website

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

Try this prompt

GPT-5

GPT-5

Openai · Aug 2025website

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

Try this prompt

Vote

9 / 38

09

Mario Level UI Recreationweb design

Recreate an interactive, classic Mario level in a single HTML file.

Google: Gemma 4 31B

Google: Gemma 4 31B

Google · Apr 2026website

Try this prompt

GPT-5

GPT-5

Openai · Aug 2025website

Try this prompt

Vote

10 / 38

10

Linear App Cloneweb design

Tests an AI's ability to replicate an existing UI with Tailwind CSS

Google: Gemma 4 31B

Google: Gemma 4 31B

Google · Apr 2026website

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

Try this prompt

GPT-5

GPT-5

Openai · Aug 2025website

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

Try this prompt

Vote

11 / 38

11

Dark Mode Dashboardweb design

Tests an AI's UI design skills with theming support

Google: Gemma 4 31B

Google: Gemma 4 31B

Google · Apr 2026website

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

Try this prompt

GPT-5

GPT-5

Openai · Aug 2025website

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

Try this prompt

Vote

12 / 38

12

Interactive Catan Boardweb design

Tests an AI's ability to create interactive web elements

Google: Gemma 4 31B

Google: Gemma 4 31B

Google · Apr 2026website

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

Try this prompt

GPT-5

GPT-5

Openai · Aug 2025website

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

Try this prompt

Vote

This matchup has more rounds

26+ 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