List your company

Hallucination Detection on BRACE Hallucination 1.0 (test)

98.8AudioCaps ScoreAQAScore
21.850441.827761.80581.7823Jan 21, 2026Jan 30, 2026Feb 9, 2026Feb 19, 2026Feb 28, 2026Mar 10, 2026Mar 20, 2026

Evaluation results

MethodLinks
AQAScoreModel=AF3-Think2026.0198.898—
AQAScoreModel=Qwen2.5-Omni-7B2026.0198.798—
FLEURBackbone=Qwen3-Omni-Instruct, Metric Architecture=LALM-ONLY2026.0398.489898.18
AQAScoreModel=Qwen2.5-Omni-3B2026.0198.397.1—
CAF-ScoreBackbone LALM=Qwen3-Omni-Instruct, Backbone CLAP=M2D-CLAP, Metric Architecture=CLAP + LALM2026.0397.9696.5997.11
CAF-ScoreBackbone LALM=Qwen3-Omni-Instruct, Backbone CLAP=MGA-CLAP, Metric Architecture=CLAP + LALM2026.0397.5796.897.09
PromptingModel=Qwen2.5-Omni-7B2026.0197.297.7—
AQAScoreModel=AF32026.0196.996.9—
FENSEReference Protocol=Reference-based2026.0396.7696.1896.4
CAF-ScoreBackbone LALM=Qwen3-Omni-Instruct, Backbone CLAP=LAION-CLAP, Metric Architecture=CLAP + LALM2026.0396.2494.9495.43
CAF-ScoreBackbone LALM=AudioFlamingo3-Base, Backbone CLAP=MGA-CLAP, Metric Architecture=CLAP + LALM2026.0395.8493.294.2
CAF-ScoreBackbone LALM=AudioFlamingo3-Base, Backbone CLAP=M2D-CLAP, Metric Architecture=CLAP + LALM2026.0395.6992.3893.63
CascadeModel=Qwen2.5-Omni-7B2026.0195.495.9—
PromptingModel=Qwen2.5-Omni-3B2026.0194.995.4—
AQAScoreModel=AF3-Chat2026.0194.694.5—
FLEURBackbone=AudioFlamingo3-Base, Metric Architecture=LALM-ONLY2026.0393.8395.8695.09
CAF-ScoreBackbone LALM=Qwen3-Omni-Instruct, Backbone CLAP=MS-CLAP, Metric Architecture=CLAP + LALM2026.0393.5296.6195.44
CAF-ScoreBackbone LALM=AudioFlamingo3-Base, Backbone CLAP=LAION-CLAP, Metric Architecture=CLAP + LALM2026.0393.289.9891.2
CLAPScoreBackbone=MGA-CLAP, Metric Architecture=CLAP-ONLY2026.0392.1887.2689.12
S-CLAPScoreBackbone=MGA-CLAP, Metric Architecture=CLAP-ONLY2026.0392.1888.0889.48
CLAPScoreModel=630k-best2026.0191.882.2—
CLAIR-AReference Protocol=Reference-based2026.0391.4591.5591.51
CascadeModel=Qwen2.5-Omni-3B2026.0191.290.2—
CLAPScoreBackbone=M2D-CLAP, Metric Architecture=CLAP-ONLY2026.0391.0383.9386.61
S-CLAPScoreBackbone=M2D-CLAP, Metric Architecture=CLAP-ONLY2026.0391.0384.4786.98
CLAPScoreModel=Music+AS2026.0190.882—
CascadeModel=AF3-Chat2026.0190.585.1—
CLAPScoreModel=Music+Speech+AS2026.019081.7—
FLEURBackbone=Qwen3-Omni-Thinking, Metric Architecture=LALM-ONLY2026.0389.786.9988.01
CLAPScoreModel=Music+Speech2026.0188.781.7—
CAF-ScoreBackbone LALM=AudioFlamingo3-Base, Backbone CLAP=MS-CLAP, Metric Architecture=CLAP + LALM2026.0388.1692.9891.16
CLAPScoreModel=630k-AS2026.0187.678—
CascadeModel=AF32026.0187.585—
PromptingModel=AF3-Chat2026.0187.185.6—
CLAPScoreBackbone=LAION-CLAP, Metric Architecture=CLAP-ONLY2026.0387.0384.3185.34
S-CLAPScoreBackbone=LAION-CLAP, Metric Architecture=CLAP-ONLY2026.0387.0384.4585.12
FLEURBackbone=AudioFlamingo3-Think, Metric Architecture=LALM-ONLY2026.0385.890.1488.5
PromptingModel=AF32026.0182.383—
Raw ScoreBackbone=Qwen3-Omni-Thinking, Metric Architecture=LALM-ONLY2026.0381.8675.878.09
Raw ScoreBackbone=Qwen3-Omni-Instruct, Metric Architecture=LALM-ONLY2026.0381.7882.4782.21
S-CLAPScoreBackbone=MS-CLAP, Metric Architecture=CLAP-ONLY2026.0373.7387.0782.04
CLAPScoreBackbone=MS-CLAP, Metric Architecture=CLAP-ONLY2026.0372.9985.2380.61
CascadeModel=AF3-Think2026.0167.574.2—
PromptingModel=AF3-Think2026.0161.285.1—
Raw ScoreBackbone=AudioFlamingo3-Base, Metric Architecture=LALM-ONLY2026.0325.2860.8347.41
Raw ScoreBackbone=AudioFlamingo3-Think, Metric Architecture=LALM-ONLY2026.0324.8121.923