Unverified50% confidenceFactExact time
Tests on the CALVIN benchmark show that when leading image editors on image-editing quality leaderboards (such as Gemini 3 Pro Image and GPT Image 1.5) are used as control targets, their success rates are lower than those of fine-tuned open-source models.
1
Sources
50%
Confidence
Long-term
Relevance
10/20/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe image editing model in the Gemini app accepts reference images plus natural language instructions and outperforms the other tools the author tested at preserving identity and lighting after editing.72% similarUnverifiedThe study's conclusions are based on the specific model Gemini 3.8 Flash and closed-ended evaluation tasks; whether selectively retrieving partial sections of profiles or facing open-ended scientific tasks would change the conclusions remains to be verified.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1074618API
curl https://kongchang.com/api/v1/knowledge/claims/1074618MCP
get_claim(id=1074618)