← All visual benchmarksArena / 2026-10-20
Arena · MULTIMODAL

Image editing

Preference for image edits.

Hover or tap a bar to inspect. Select up to 3 models to highlight and compare.
Model scores10 models
Image editing. Zoomed horizontal axis, from 1380.5 to 1531.9ZOOMED X-AXIS · 1380.5 – 1531.9 · NOT FROM ZERO1380.51418.41456.21494.01531.9GPT-Image-2.5 SunburstGPT-Image-2.5 Sunburst · 1514.4 Arena Rating · zoomed horizontal axis1514.4GPT-Image-2.5 FlareGPT-Image-2.5 Flare · 1480.7 Arena Rating · zoomed horizontal axis1480.7GPT-Image-2 (Medium)GPT-Image-2 (Medium) · 1460.6 Arena Rating · zoomed horizontal axis1460.6Grok Imagine Image 2.0Grok Imagine Image 2.0 · 1439.4 Arena Rating · zoomed horizontal axis1439.4Gemini Nano Banana 2.1Gemini Nano Banana 2.1 · 1426.2 Arena Rating · zoomed horizontal axis1426.2MAI-Image-2.6MAI-Image-2.6 · 1425.3 Arena Rating · zoomed horizontal axis1425.3Grok Imagine Image 2.0 (20260801)Grok Imagine Image 2.0 (20260801) · 1415 Arena Rating · zoomed horizontal axis1415Krea 3 UltraKrea 3 Ultra · 1410.5 Arena Rating · zoomed horizontal axis1410.5MAI-Image-2.5MAI-Image-2.5 · 1399.9 Arena Rating · zoomed horizontal axis1399.9Muse ImageMuse Image · 1398 Arena Rating · zoomed horizontal axis1398
Choose a bar for details.

How to read this chart

Official top-ranked excerpt. Human preference ratings are not accuracy percentages; IPS is a distinct metric.

Score unit: Arena Rating. Values can only be compared within the same task and protocol.

Source publication: 2026-10-20. Higher values indicate better results for the reported metric.

Source: Arena · CC BY 4.0. Redrawn from released data; reduced to selected models and rounded. Not affiliated with the source organization.

Explore related charts