Updated Feb 6, 2026

Our Verdict

Qwen3 235B A22B

Pony Alpha

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

Too close to call

Writing DNA

Style Comparison

Similarity

96%

Pony Alpha uses 40.4x more hedging

Qwen3 235B A22B

Pony Alpha

68%Vocabulary52%

9wSentence Length15w

0.00Hedging0.40

5.8Bold5.9

3.7Lists4.4

0.74Emoji0.00

0.00Headings0.71

0.12Transitions0.23

Based on 2 + 20 text responses

vs

Ask them anything yourself

Qwen3 235B A22B

Pony Alpha

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

Qwen3 235B A22B vs MiniMax M2.5New provider

Qwen3 235B A22B vs GPT OSS 20BNew provider

Qwen3 235B A22B vs Claude Opus 4New provider

Qwen3 235B A22BvsPony Alpha

47 fights queued

Why Qwen3 235B A22B?

Dead even. This one's a coin flip.

Why Pony Alpha?

6.1x more context200K

9 months newerFeb 2026

Qwen3 235B A22BPony Alpha

Input price

—

$0.000/M

Output price

—

$0.000/M

Context

33K

200K

Released

Apr 2025

Feb 2026

Favorites

Qwen3 235B A22BPony Alpha

Movie

The Godfather

The Godfather

—

Album

Sgt Peppers Lonely Hearts Club Band

Sgt Peppers Lonely Hearts Club Band

—

Book

Nineteen Eighty-Four

Nineteen Eighty-Four

—

City

Istanbul

—

Game

The Legend of Zelda: Ocarina of Time

The Legend of Zelda: Ocarina of Time

Adventure, Action4.4

—

12 / 47

01

Estimate Complexityreasoning

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

Qwen3 235B A22B

Qwen3 235B A22B

Qwen · Apr 2025text

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

Try this prompt

Pony Alpha

Pony Alpha

Openrouter · Feb 2026text

The total number of FLOPs required to train GPT‑3 is approximately 3.15 × 10²³ (or 315 zettaFLOPs). The reasoning follows these steps:

Model size – GPT‑3 has 175 billion (175 × 10⁹) parameters. This is a well‑known figure from the original paper.
Training data size – The model was trained on 300 billion (300 × 10⁹) tokens. The paper states that all models were trained for a total of 300 billion tokens.
FLOPs per parameter per token – For a transformer model, the compute can be approximated by:
- Forward pass: each parameter is used in a multiply‑add operation, contributing about 2 FLOPs per token.
- Backward pass (gradient computation): roughly 4 FLOPs per parameter per token (two extra passes for gradients with respect to weights and inputs).
- Hence, training (forward + backward) requires about 6 FLOPs per parameter per token.
Total FLOPs – Multiply the three numbers: [ \text{FLOPs} = 6 \times (175 \times 10^9) \times (300 \times 10^9) = 6 \times 175 \times 300 \times 10^{18} = 315,!000 \times 10^{18} = 3.15 \times 10^{23}. ]
Consistency check – This estimate aligns with the reported compute of 3.14 × 10²³ FLOPs mentioned in OpenAI’s technical analysis, confirming its plausibility.

The calculation captures the dominant linear‑algebra operations; small additional costs (e.g., attention quadratic terms, layer norms) are negligible in comparison.

Try this prompt

Vote

2 / 47

02

AI Board Game Logicreasoning

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

Qwen3 235B A22B

Qwen3 235B A22B

Qwen · Apr 2025svg

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

Try this prompt

Pony Alpha

Pony Alpha

Openrouter · Feb 2026svg

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

Try this prompt

Vote

3 / 47

03

Logic Puzzlereasoning

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

Qwen3 235B A22B

Qwen3 235B A22B

Qwen · Apr 2025text

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

Try this prompt

Pony Alpha

Pony Alpha

Openrouter · Feb 2026text

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

Try this prompt

Vote

4 / 47

04

Stochastic Consistency Testanalysis

Tests an AI's randomness and creativity

Qwen3 235B A22B

Qwen3 235B A22B

Qwen · Apr 2025text

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

Try this prompt

Pony Alpha

Pony Alpha

Openrouter · Feb 2026text

Here are 5 jokes:

Why don't scientists trust atoms? Because they make up everything!
I told my wife she was drawing her eyebrows too high. She looked surprised.
What do you call a fake noodle? An impasta.
Why did the scarecrow win an award? Because he was outstanding in his field.
I'm reading a book about anti-gravity. It's impossible to put down.

There are 5 unique jokes.

Try this prompt

Vote

5 / 47

05

SVG Layout Challengeimage generation

Tests an AI's ability to generate vector graphics

Qwen3 235B A22B

Qwen3 235B A22B

Qwen · Apr 2025svg

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

Try this prompt

Pony Alpha

Pony Alpha

Openrouter · Feb 2026svg

Try this prompt

Vote

6 / 47

06

Xbox Controller SVG Artimage generation

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

Qwen3 235B A22B

Qwen3 235B A22B

Qwen · Apr 2025svg

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

Try this prompt

Pony Alpha

Pony Alpha

Openrouter · Feb 2026svg

Try this prompt

Vote

Sponsored

7 / 47

07

Generate a Stand-Up Routineconversation

Tests an AI's humor and creative writing ability

Qwen3 235B A22B

Qwen3 235B A22B

Qwen · Apr 2025text

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

Try this prompt

Pony Alpha

Pony Alpha

Openrouter · Feb 2026text

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

Try this prompt

Vote

8 / 47

08

Realistic AI Interviewconversation

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

Qwen3 235B A22B

Qwen3 235B A22B

Qwen · Apr 2025text

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

Try this prompt

Pony Alpha

Pony Alpha

Openrouter · Feb 2026text

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

Try this prompt

Vote

9 / 47

09

Character Voice Testconversation

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

Qwen3 235B A22B

Qwen3 235B A22B

Qwen · Apr 2025text

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

Try this prompt

Pony Alpha

Pony Alpha

Openrouter · Feb 2026text

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

Try this prompt

Vote

10 / 47

10

Minimalist Landing Pageweb design

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

Qwen3 235B A22B

Qwen3 235B A22B

Qwen · Apr 2025website

Try this prompt

Pony Alpha

Pony Alpha

Openrouter · Feb 2026website

Try this prompt

Vote

11 / 47

11

Pokémon Battle UI Recreationweb design

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

Qwen3 235B A22B

Qwen3 235B A22B

Qwen · Apr 2025website

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

Try this prompt

Pony Alpha

Pony Alpha

Openrouter · Feb 2026website

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

Try this prompt

Vote

12 / 47

12

Three.js 3D Gameweb design

Create a 3D building and driving game in a single self-contained HTML file using Three.js. Tests an AI's ability to combine 3D graphics, game logic, and UI interactions on the web.

Qwen3 235B A22B

Qwen3 235B A22B

Qwen · Apr 2025html

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

Try this prompt

Pony Alpha

Pony Alpha

Openrouter · Feb 2026html

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

Try this prompt

Vote

This matchup has more rounds

35+ 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