Grok 4.6 vs Opus 5
The same one-shot prompt, one attempt each, and no fixes. I chose the winner after playing both games.
VIEW TEST ON X- Prompt
- Identical
- Attempts
- One each
- Fixes
- Zero
- Winner
- Grok 4.6
TESTS / 02 ENTRIES
Model tests without the mystery.
Same inputs when possible. Attempts and fixes disclosed. The original post linked directly.
mona-lisa-1 vs GPT Image
A mystery Arena model and GPT Image received the same prompt. The outputs were shown side by side.
VIEW TEST ON X- Prompt
- Identical
- Format
- Side by side
- Model
- mona-lisa-1
- Compared with
- GPT Image
THE METHOD
Four rules for every comparison.
- 01Use the same inputs when possible.
- 02Disclose attempts, edits, and settings.
- 03Show both outputs.
- 04Separate facts from personal taste.