Volumetric Computed Tomography Multiple Choice Question Answering on CT-RATE MCQ (test)
59.96Existence AccuracyTIF-GRPO w/ Acc Reward (CT-RATE)
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| TIF-GRPO w/ Acc Reward (CT-RATE)#Params=4B2026.05 | 59.96 | 73.84 | 89.37 | 70.25 | 66.65 | |
| GRPO w/ Acc Reward (CT-RATE)#Params=4B2026.05 | 55.03 | 73 | 88.41 | 67.29 | 63.59 | |
| TIF-GRPO w/ Acc Reward (AMOS-MM)#Params=4B2026.05 | 55.03 | 68.35 | 85.99 | 65.53 | 69.67 | |
| GRPO w/ Acc Reward (AMOS-MM)#Params=4B2026.05 | 54.18 | 61.18 | 73.91 | 60.48 | 63.91 | |
| GPT-5#Params=*2026.05 | 53.53 | 56.96 | 83.57 | 61.25 | 60.54 | |
| Qwen3-VL-8B#Params=8B2026.05 | 52.89 | 46.84 | 61.35 | 53.24 | 50.63 | |
| Fleming-VL-8B#Params=8B2026.05 | 52.89 | 55.27 | 65.7 | 56.42 | 53.92 | |
| Hulu-Med#Params=7B2026.05 | 52.18 | 58.76 | 81.09 | 60.46 | 57.03 | |
| RadFM#Params=14B2026.05 | 51.82 | 28.69 | 26.09 | 39.96 | 39.53 | |
| InternVL3.5-8B#Params=8B2026.05 | 51.18 | 45.15 | 63.77 | 52.47 | 49.18 | |
| CT-CHAT#Params=8B2026.05 | 49.25 | 55.27 | 66.67 | 54.77 | 54.22 | |
| M3D-LaMed#Params=7B2026.05 | 47.54 | 53.16 | 63.29 | 52.58 | 48.72 | |
| Gemini-3-Pro#Params=*2026.05 | 46.89 | 56.54 | 80.19 | 56.97 | 56.7 |