There are no watermarks in the arena.

There are no visible watermarks, but model makers can use steganographic codes to identify outputs from their own models.

Text-to-Image Models Leave Identifiable Signatures: Implications for Leaderboard Security

	▲	encroach 5 days ago \| parent [-]
		This is true, however LMArena does employ some methods to mitigate attempts to manipulate the leaderboard, see https://openreview.net/forum?id=zf9zwCRKyP They also control for style https://news.lmarena.ai/sentiment-control/