Google
·
RELEASED
Imagen 3.0 Generate 002
Text-to-Image
of 62 ranked
4.0 of 5
Especially good for Photorealism and Prompt Adherence.
Standout work
The model at its best across 4 more challenges.
Skills
The shape of this model's strengths, and the categories it ranks strongest in.
Where Imagen 3.0 Generate 002 earns its score, percentile-shaped across its skills.
-
1
Prompt Adherence 23.1
-
2
Photorealism 23.1
-
3
Art 21.1
-
4
Product, Branding & Commercial 20.8
-
5
Aesthetics 19.9
By skill
Every challenge Imagen 3.0 Generate 002 entered, grouped by the skill it tests and led by its strongest lanes. Scroll a lane to see them all; open a stack to compare its takes.
Show 5 more skill lanes
Notable matchups
This model's standout head-to-heads from the arena's blind voting. Open one for the full comparison, or stack it up against any other model.
Put Imagen 3.0 Generate 002 up against another model.
61 compatible models to stack it up against.