List your company

Complex Instruction Following on InstructTTSEval-ZH

90.9APSGround-truth
45.76457.48269.280.918Jan 8, 2026Feb 1, 2026Feb 26, 2026Mar 23, 2026Apr 16, 2026May 11, 2026Jun 5, 2026

Evaluation results

MethodLinks
Ground-truthSource Type=Baseline2026.0190.986.769.882.5
Gemini-proSource Type=Closed-sourced2026.018990.175.584.8
Gemini-TTS-ProVoice Control Type=fixed voice editing2026.038990.175.5—
Gemini-TTS-Pro2026.068990.175.5—
Gemini-flashSource Type=Closed-sourced2026.0188.290.977.385.4
Qwen3-TTS-12Hz-1.7B-VDFrequency=12Hz, Parameters=1.7B2026.0685.281.165.1—
VoxCPM22026.0685.271.560.8—
Qwen3-TTS-VD2026.0384.382.977.4—
MOSS-VoiceGenerator2026.03788074—
MOSS-VoiceGenerator2026.06788074—
MiMo-Audio-7B-InstructSource Type=Open-sourced2026.0175.774.361.570.5
VoiceSculptor-VD2026.0375.764.761.5—
MIMO-Audio-7B-Instruct2026.0375.774.361.5—
VoiceSculptor2026.0675.764.761.5—
Mimo-Audio-7B-InstructParameters=7B2026.0675.774.361.5—
FlexiVoiceSource Type=Open-sourced2026.017171.869.770.8
FlexiVoice-BaseSource Type=Open-sourced2026.0156.759.159.558.4
GPT-4o-mini-TTSSource Type=Closed-sourced2026.0154.952.34651.1
GPT-4o-mini-TTSVoice Control Type=fixed voice editing2026.0354.952.346—
GPT-4o-mini-TTS2026.0654.952.346—
VoxInstructSource Type=Open-sourced2026.0147.552.342.647.5
VoxInstructRequires reference audio=true2026.0347.552.342.6—
VoxInstruct2026.0647.552.342.6—