Radiology Question Answering on RadQA and Board-RadQA Combined (test)
91AccuracyGemini-3.1-Pro
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Gemini-3.1-ProPrompting Strategy=Agentic2026.03 | 91 | 154 | 0.012 | |
| DeepSeek-V3Prompting Strategy=Agentic2026.03 | 89 | 150 | 0.218 | |
| o3Prompting Strategy=Agentic2026.03 | 89 | 150 | 0.239 | |
| Claude-Sonnet-4.6Prompting Strategy=Agentic2026.03 | 88 | 148 | 0.577 | |
| DeepSeek-R1Prompting Strategy=Zero-shot2026.03 | 88 | 148 | — | |
| MiMo-V2-FlashPrompting Strategy=Agentic2026.03 | 87 | 147 | 0.063 | |
| GPT-5Prompting Strategy=Agentic2026.03 | 87 | 147 | 0.402 | |
| GPT-5.2Prompting Strategy=Agentic2026.03 | 87 | 147 | 0.035 | |
| Palmyra-X5Prompting Strategy=Zero-shot2026.03 | 86 | 146 | — | |
| Palmyra-X5Prompting Strategy=Agentic2026.03 | 86 | 145 | 1 | |
| Llama3.3-70BPrompting Strategy=Agentic2026.03 | 86 | 145 | 0.144 | |
| Mistral Large (123B)Prompting Strategy=Agentic2026.03 | 86 | 146 | 0.166 | |
| Qwen3-235BPrompting Strategy=Zero-shot2026.03 | 86 | 146 | — | |
| Claude-Sonnet-4.6Prompting Strategy=Zero-shot2026.03 | 85 | 144 | — | |
| GLM-5Prompting Strategy=Agentic2026.03 | 85 | 144 | 0.527 | |
| Kimi-K2.5Prompting Strategy=Agentic2026.03 | 85 | 144 | 0.729 | |
| Llama4-Scout-16EPrompting Strategy=Agentic2026.03 | 85 | 143 | 0.585 | |
| o3Prompting Strategy=Zero-shot2026.03 | 85 | 144 | — | |
| Qwen3-235BPrompting Strategy=Agentic2026.03 | 85 | 144 | 0.877 | |
| DeepSeek R1-70BPrompting Strategy=Zero-shot2026.03 | 84 | 142 | — | |
| DeepSeek R1-70BPrompting Strategy=Agentic2026.03 | 84 | 142 | 1 | |
| DeepSeek-R1Prompting Strategy=Agentic2026.03 | 84 | 142 | 0.343 | |
| DeepSeek-V3Prompting Strategy=Zero-shot2026.03 | 84 | 142 | — | |
| GPT-5Prompting Strategy=Zero-shot2026.03 | 84 | 142 | — | |
| MedGemma-27B-text-itPrompting Strategy=Agentic2026.03 | 84 | 142 | 0.22 | |
| Qwen2.5-70BPrompting Strategy=Agentic2026.03 | 84 | 142 | 0.229 | |
| Kimi-K2.5Prompting Strategy=Zero-shot2026.03 | 83 | 141 | — | |
| Gemma-3-27B-itPrompting Strategy=Agentic2026.03 | 83 | 141 | 0.009 | |
| Gemini-3.1-ProPrompting Strategy=Zero-shot2026.03 | 82 | 138 | — | |
| GLM-5Prompting Strategy=Zero-shot2026.03 | 82 | 139 | — | |
| Llama4-Scout-16EPrompting Strategy=Zero-shot2026.03 | 82 | 139 | — | |
| GPT-4-turboPrompting Strategy=Agentic2026.03 | 82 | 138 | 0.277 | |
| Mistral Large (123B)Prompting Strategy=Zero-shot2026.03 | 81 | 137 | — | |
| Qwen3-8BPrompting Strategy=Agentic2026.03 | 81 | 137 | 0.218 | |
| MiniMax-M2.5Prompting Strategy=Agentic2026.03 | 80 | 135 | 0.999 | |
| Llama3.3-70BPrompting Strategy=Zero-shot2026.03 | 80 | 135 | — | |
| MiniMax-M2.5Prompting Strategy=Zero-shot2026.03 | 79 | 134 | — | |
| MiMo-V2-FlashPrompting Strategy=Zero-shot2026.03 | 79 | 133 | — | |
| Llama3-Med42-70BPrompting Strategy=Agentic2026.03 | 79 | 134 | 0.417 | |
| GPT-5.2Prompting Strategy=Zero-shot2026.03 | 79 | 134 | — | |
| MedGemma-27B-text-itPrompting Strategy=Zero-shot2026.03 | 79 | 133 | — | |
| Qwen2.5-14BPrompting Strategy=Agentic2026.03 | 79 | 134 | 0.354 | |
| Qwen2.5-70BPrompting Strategy=Zero-shot2026.03 | 79 | 134 | — | |
| Qwen2.5-7BPrompting Strategy=Agentic2026.03 | 78 | 132 | 0.002 | |
| GPT-3.5-turboPrompting Strategy=Agentic2026.03 | 77 | 130 | 0.012 | |
| GPT-4-turboPrompting Strategy=Zero-shot2026.03 | 77 | 130 | — | |
| LFM2.5-1.2B-ThinkingPrompting Strategy=Agentic2026.03 | 76 | 129 | 0.001 | |
| Ministral-8BPrompting Strategy=Agentic2026.03 | 76 | 128 | 0.001 | |
| Llama3.3-8BPrompting Strategy=Agentic2026.03 | 75 | 126 | 0.275 | |
| Llama3-Med42-70BPrompting Strategy=Zero-shot2026.03 | 75 | 127 | — | |
| MedGemma-4B-itPrompting Strategy=Agentic2026.03 | 75 | 127 | 0.024 | |
| Qwen3-8BPrompting Strategy=Zero-shot2026.03 | 75 | 127 | — | |
| Llama3-Med42-8BPrompting Strategy=Agentic2026.03 | 74 | 125 | 0.024 | |
| Qwen2.5-14BPrompting Strategy=Zero-shot2026.03 | 74 | 126 | — | |
| Qwen2.5-3BPrompting Strategy=Agentic2026.03 | 73 | 124 | 0.038 | |
| Gemma-3-4B-itPrompting Strategy=Agentic2026.03 | 71 | 121 | 0.001 | |
| Gemma-3-27B-itPrompting Strategy=Zero-shot2026.03 | 71 | 120 | — | |
| Llama3.3-8BPrompting Strategy=Zero-shot2026.03 | 69 | 117 | — | |
| GPT-3.5-turboPrompting Strategy=Zero-shot2026.03 | 64 | 108 | — | |
| MedGemma-4B-itPrompting Strategy=Zero-shot2026.03 | 64 | 108 | — | |
| Llama3-Med42-8BPrompting Strategy=Zero-shot2026.03 | 63 | 107 | — | |
| Qwen2.5-3BPrompting Strategy=Zero-shot2026.03 | 63 | 107 | — | |
| Qwen2.5-7BPrompting Strategy=Zero-shot2026.03 | 63 | 107 | — | |
| LFM2.5-1.2B-ThinkingPrompting Strategy=Zero-shot2026.03 | 56 | 94 | — | |
| Ministral-8BPrompting Strategy=Zero-shot2026.03 | 54 | 91 | — | |
| Gemma-3-4B-itPrompting Strategy=Zero-shot2026.03 | 51 | 86 | — | |
| Qwen2.5-0.5BPrompting Strategy=Agentic2026.03 | 48 | 81 | 0.079 | |
| Qwen2.5-0.5BPrompting Strategy=Zero-shot2026.03 | 37 | 63 | — |