# design-audiorealism benchmark

| \# | model | design-audiorealism | $/task | card | cheapest list | cheapest effective | unit | providers | tokens/week |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | `microsoft/mai-voice-2` | 1208.0 |  |  | 22.0 azure |  | $0.000022/M characters Characters azure | 1 | 151886 |
| 2 | `microsoft/mai-voice-2-flash` | 1137.0 |  |  | 15.0 azure |  | $0.000015/M characters Characters azure | 1 | 147625 |
| 3 | `google/gemini-3.8-flash-tts` | 1080.0 |  |  | 0.5 google-ai-studio | 0.4915 google-ai-studio |  | 2 | 15001320 |
| 4 | `fish-audio/drama-3` | 958.0 |  | 10.0 | 10.0 fish-audio |  | $0.00001/M UTF-8 bytes UTF-8 Bytes fish-audio | 1 | 1686200 |
| 5 | `fish-audio/s2.1-pro` | 953.0 |  | 15.0 | 15.0 fish-audio |  | $0.000015/M UTF-8 bytes UTF-8 Bytes fish-audio | 1 | 215178 |
| 6 | `qwen/qwen-audio-3.0-tts-flash` | 854.0 |  |  | 15.0 alibaba |  | $0.000015/M characters Characters alibaba | 1 | 40623 |