AA-Music-Vocal v1.1Artificial Analysis
Range | Model | Elo | Samples | |
|---|---|---|---|---|
| 1 | 1-2 |  ACE Music v1 | 1160±2195% interval 1139 to 1181 | 2,119 |
| 2 | 1-3 |  Suno v6 | 1144±1895% interval 1126 to 1162 | 3,063 |
| 3 | 2-3 | Suno v6-mini | 1132±1895% interval 1114 to 1150 | 3,175 |
| 4 | 4 |  Mureka V9.5 | 1101±1895% interval 1083 to 1119 | 3,141 |
| 5 | 5-6 | Mureka V9 | 1063±1895% interval 1045 to 1081 | 3,121 |
| 6 | 5-7 |  Lyria 3.5 | 1050±1895% interval 1032 to 1068 | 3,203 |
| 7 | 5-7 | Lyria 3 Pro | 1048±1895% interval 1030 to 1066 | 3,191 |
| 8 | 8-10 |  Eleven Music v2.5 | 1016±1895% interval 998 to 1034 | 3,248 |
| 9 | 8-13 | Eleven Music v2 | 1002±1895% interval 984 to 1020 | 3,295 |
| 10 | 8-13 |  MiniMax Music 3.0 | 1000AnchorFixed at 1000 as the reference point for this board's Elo scale. | 3,336 |
| 11 | 9-13 | Eleven Music | 998±1895% interval 980 to 1016 | 3,246 |
| 12 | 9-13 | MiniMax Music 2.6 | 998±1895% interval 980 to 1016 | 3,231 |
| 13 | 10-13 | MiniMax Music 2.5+ | 990±1895% interval 972 to 1008 | 3,362 |
| 14 | 14 |  Melodia v3 | 964±1895% interval 946 to 982 | 3,319 |
Frequently Asked Questions
ACE Music v1 leads AA-Music-Vocal v1.1 with an Elo score of 1160.
The top models on AA-Music-Vocal v1.1 by Elo rating are: 1. ACE Music v1 (Elo 1160), 2. Suno v6 (Elo 1144), 3. Suno v6-mini (Elo 1132), 4. Mureka V9.5 (Elo 1101), and 5. Mureka V9 (Elo 1063). Rankings are based on blind votes from our validated private evaluators.
Models are ranked using an Elo rating system derived from blind comparisons judged by our validated private evaluators. Evaluators listen to pairs of music samples generated from the same prompt and choose their preferred track without knowing which model created it. Higher Elo scores indicate a model is preferred more often. Vote in the Artificial Analysis Music Arena
AA-Music-Instrumental v1.1 ranks models generating music without vocals, while AA-Music-Vocal v1.1 ranks models that generate music with singing or vocal elements. Models may perform differently in each benchmark, so separate rankings are provided.