We're launching AA-Music-Vocal v1.1 and AA-Music-Instrumental v1.1, the new benchmarks behind the Artificial Analysis Music Arena, built on 1,000 new prompts across 17 genres and scored by our recruited evaluator panel.
In the last 3 months, flagship model releases like Suno v6, Lyria 3.5, Mureka V9.5, Eleven Music v2.5 and MiniMax Music 3.0 have set new bars for quality in AI music generation. Models now write complete songs from a single text prompt, producing intricate lyrics, convincing vocals, complex arrangements and authentic instrumentation for the genre. As quality rises, the differences between models get subtler, demanding better and more nuanced evaluations for our AA-Music leaderboards.
Initial insights from the updated Artificial Analysis Music Leaderboards:
➤ Suno v6 ranks #1 on both benchmarks, 26 Elo ahead of Suno v6-mini on Vocal and 31 Elo ahead on Instrumental. Suno holds the top two places on both leaderboards.
➤ Mureka V9.5 ranks #3 on Vocal, 51 Elo above Mureka V9, the largest gain between two versions of the same model family on the Vocal leaderboard.
➤ Mureka V9, Mureka V9.5, Lyria 3 Pro and Lyria 3.5 are statistically tied at #3 to #6 on Instrumental, within 5 Elo of each other.
➤ MiniMax Music 3.0 is the leading open weights music model, at #11 on Vocal and #13 on Instrumental. Stable Audio 3 Medium is within 3 Elo of it on Instrumental.
See below for what's new in v1.1 and example tracks from leading models🧵