List your company

Visual Question Answering on VQA-RAD (Open)

0.819AUROCLLaVA-NeXT-34B
0.42380.52640.6290.7316Apr 2, 2026

Evaluation results

MethodLinks
LLaVA-NeXT-34BCalibration Method=HAC-Gate, Confidence Type=Verbalized2026.040.819
InternVL3-2BCalibration Method=HAC-Gate, Confidence Type=Sampling2026.040.815
InternVL3-2BCalibration Method=HAC-Gate, Confidence Type=Verbalized2026.040.808
LLaVA-NeXT-34BCalibration Method=HAC-Platt, Confidence Type=Verbalized2026.040.807
InternVL3-2BCalibration Method=HAC-Platt, Confidence Type=Verbalized2026.040.806
InternVL3-2BCalibration Method=HAC-Platt, Confidence Type=Sampling2026.040.803
Qwen3-VL-32BCalibration Method=HAC-Gate, Confidence Type=Verbalized2026.040.798
InternVL3-38BCalibration Method=HAC-Platt, Confidence Type=Sampling2026.040.796
InternVL3-38BCalibration Method=HAC-Gate, Confidence Type=Sampling2026.040.794
Qwen3-VL-32BCalibration Method=HAC-Platt, Confidence Type=Verbalized2026.040.788
LLaVA-NeXT-7BCalibration Method=HAC-Platt, Confidence Type=Sampling2026.040.788
LLaVA-NeXT-7BCalibration Method=HAC-Gate, Confidence Type=Sampling2026.040.788
Qwen3-VL-8BCalibration Method=HAC-Platt, Confidence Type=Verbalized2026.040.774
InternVL3-38BCalibration Method=HAC-Platt, Confidence Type=Verbalized2026.040.771
InternVL3-38BCalibration Method=Uncalibrated, Confidence Type=Sampling2026.040.769
InternVL3-38BCalibration Method=Platt Scaling, Confidence Type=Sampling2026.040.769
InternVL3-38BCalibration Method=Isotonic Regr., Confidence Type=Sampling2026.040.768
Qwen3-VL-8BCalibration Method=Uncalibrated, Confidence Type=Verbalized2026.040.753
Qwen3-VL-8BCalibration Method=Platt Scaling, Confidence Type=Verbalized2026.040.753
Qwen3-VL-8BCalibration Method=Isotonic Regr., Confidence Type=Verbalized2026.040.753
Qwen3-VL-32BCalibration Method=HAC-Platt, Confidence Type=Sampling2026.040.753
LLaVA-NeXT-7BCalibration Method=Isotonic Regr., Confidence Type=Sampling2026.040.749
Qwen3-VL-8BCalibration Method=HAC-Gate, Confidence Type=Verbalized2026.040.746
Qwen3-VL-32BCalibration Method=HAC-Gate, Confidence Type=Sampling2026.040.745
LLaVA-NeXT-7BCalibration Method=Uncalibrated, Confidence Type=Sampling2026.040.745
LLaVA-NeXT-7BCalibration Method=Platt Scaling, Confidence Type=Sampling2026.040.745
InternVL3-8BCalibration Method=HAC-Platt, Confidence Type=Verbalized2026.040.739
InternVL3-38BCalibration Method=Hist. Binning, Confidence Type=Sampling2026.040.738
LLaVA-NeXT-7BCalibration Method=HAC-Platt, Confidence Type=Verbalized2026.040.737
LLaVA-NeXT-7BCalibration Method=HAC-Gate, Confidence Type=Verbalized2026.040.737
Qwen3-VL-8BCalibration Method=Hist. Binning, Confidence Type=Verbalized2026.040.735
LLaVA-NeXT-34BCalibration Method=Isotonic Regr., Confidence Type=Sampling2026.040.728
InternVL3-2BCalibration Method=Uncalibrated, Confidence Type=Sampling2026.040.727
InternVL3-2BCalibration Method=Platt Scaling, Confidence Type=Sampling2026.040.727
InternVL3-38BCalibration Method=HAC-Gate, Confidence Type=Verbalized2026.040.726
InternVL3-38BCalibration Method=Uncalibrated, Confidence Type=Verbalized2026.040.72
InternVL3-38BCalibration Method=Platt Scaling, Confidence Type=Verbalized2026.040.72
Qwen3-VL-2BCalibration Method=HAC-Gate, Confidence Type=Verbalized2026.040.717
Qwen3-VL-8BCalibration Method=HAC-Platt, Confidence Type=Sampling2026.040.717
Qwen3-VL-8BCalibration Method=HAC-Gate, Confidence Type=Sampling2026.040.717
Qwen3-VL-2BCalibration Method=HAC-Platt, Confidence Type=Verbalized2026.040.716
InternVL3-38BCalibration Method=Isotonic Regr., Confidence Type=Verbalized2026.040.716
InternVL3-38BCalibration Method=Hist. Binning, Confidence Type=Verbalized2026.040.715
InternVL3-8BCalibration Method=HAC-Gate, Confidence Type=Verbalized2026.040.713
LLaVA-NeXT-34BCalibration Method=Uncalibrated, Confidence Type=Sampling2026.040.713
LLaVA-NeXT-34BCalibration Method=Platt Scaling, Confidence Type=Sampling2026.040.713
Qwen3-VL-32BCalibration Method=Uncalibrated, Confidence Type=Verbalized2026.040.708
Qwen3-VL-32BCalibration Method=Platt Scaling, Confidence Type=Verbalized2026.040.708
InternVL3-2BCalibration Method=Isotonic Regr., Confidence Type=Sampling2026.040.705
InternVL3-8BCalibration Method=HAC-Gate, Confidence Type=Sampling2026.040.704
LLaVA-NeXT-34BCalibration Method=HAC-Platt, Confidence Type=Sampling2026.040.704
LLaVA-NeXT-34BCalibration Method=Uncalibrated, Confidence Type=Verbalized2026.040.701
LLaVA-NeXT-34BCalibration Method=Platt Scaling, Confidence Type=Verbalized2026.040.701
InternVL3-8BCalibration Method=HAC-Platt, Confidence Type=Sampling2026.040.699
LLaVA-NeXT-34BCalibration Method=Isotonic Regr., Confidence Type=Verbalized2026.040.699
Qwen3-VL-32BCalibration Method=Isotonic Regr., Confidence Type=Verbalized2026.040.698
LLaVA-NeXT-34BCalibration Method=HAC-Gate, Confidence Type=Sampling2026.040.696
LLaVA-NeXT-34BCalibration Method=Hist. Binning, Confidence Type=Sampling2026.040.69
Qwen3-VL-2BCalibration Method=HAC-Gate, Confidence Type=Sampling2026.040.676
Qwen3-VL-8BCalibration Method=Isotonic Regr., Confidence Type=Sampling2026.040.676
Qwen3-VL-8BCalibration Method=Uncalibrated, Confidence Type=Sampling2026.040.675
Qwen3-VL-8BCalibration Method=Platt Scaling, Confidence Type=Sampling2026.040.675
Qwen3-VL-2BCalibration Method=HAC-Platt, Confidence Type=Sampling2026.040.672
Qwen3-VL-2BCalibration Method=Uncalibrated, Confidence Type=Sampling2026.040.666
Qwen3-VL-2BCalibration Method=Platt Scaling, Confidence Type=Sampling2026.040.666
InternVL3-2BCalibration Method=Hist. Binning, Confidence Type=Sampling2026.040.663
LLaVA-NeXT-34BCalibration Method=Hist. Binning, Confidence Type=Verbalized2026.040.662
InternVL3-8BCalibration Method=Uncalibrated, Confidence Type=Sampling2026.040.654
InternVL3-8BCalibration Method=Platt Scaling, Confidence Type=Sampling2026.040.654
Qwen3-VL-32BCalibration Method=Hist. Binning, Confidence Type=Verbalized2026.040.653
Qwen3-VL-2BCalibration Method=Isotonic Regr., Confidence Type=Sampling2026.040.65
InternVL3-8BCalibration Method=Isotonic Regr., Confidence Type=Sampling2026.040.64
LLaVA-NeXT-7BCalibration Method=Hist. Binning, Confidence Type=Sampling2026.040.631
Qwen3-VL-2BCalibration Method=Uncalibrated, Confidence Type=Verbalized2026.040.625
Qwen3-VL-2BCalibration Method=Platt Scaling, Confidence Type=Verbalized2026.040.625
InternVL3-8BCalibration Method=Isotonic Regr., Confidence Type=Verbalized2026.040.617
InternVL3-8BCalibration Method=Uncalibrated, Confidence Type=Verbalized2026.040.616
InternVL3-8BCalibration Method=Platt Scaling, Confidence Type=Verbalized2026.040.616
Qwen3-VL-8BCalibration Method=Hist. Binning, Confidence Type=Sampling2026.040.614
InternVL3-8BCalibration Method=Hist. Binning, Confidence Type=Verbalized2026.040.604
Qwen3-VL-2BCalibration Method=Hist. Binning, Confidence Type=Sampling2026.040.601
Qwen3-VL-2BCalibration Method=Isotonic Regr., Confidence Type=Verbalized2026.040.599
InternVL3-8BCalibration Method=Hist. Binning, Confidence Type=Sampling2026.040.598
LLaVA-NeXT-7BCalibration Method=Uncalibrated, Confidence Type=Verbalized2026.040.589
LLaVA-NeXT-7BCalibration Method=Platt Scaling, Confidence Type=Verbalized2026.040.589
Qwen3-VL-2BCalibration Method=Hist. Binning, Confidence Type=Verbalized2026.040.587
LLaVA-NeXT-7BCalibration Method=Isotonic Regr., Confidence Type=Verbalized2026.040.581
Qwen3-VL-32BCalibration Method=Isotonic Regr., Confidence Type=Sampling2026.040.579
Qwen3-VL-32BCalibration Method=Platt Scaling, Confidence Type=Sampling2026.040.578
Qwen3-VL-32BCalibration Method=Uncalibrated, Confidence Type=Sampling2026.040.577
Qwen3-VL-32BCalibration Method=Hist. Binning, Confidence Type=Sampling2026.040.548
LLaVA-NeXT-7BCalibration Method=Hist. Binning, Confidence Type=Verbalized2026.040.536
InternVL3-2BCalibration Method=Uncalibrated, Confidence Type=Verbalized2026.040.514
InternVL3-2BCalibration Method=Platt Scaling, Confidence Type=Verbalized2026.040.514
InternVL3-2BCalibration Method=Isotonic Regr., Confidence Type=Verbalized2026.040.468
InternVL3-2BCalibration Method=Hist. Binning, Confidence Type=Verbalized2026.040.439