Analytical Agent Reasoning on CGM-QA synthetic queries N=2,470
92PrecisionGemini 3.0 Pro
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Gemini 3.0 ProModel Accessibility=Proprietary2026.04 | 92 | 71 | 80 | 94 | |
| Gemini 3.0 FlashModel Accessibility=Proprietary2026.04 | 89 | 74 | 81 | 94 | |
| GPT-5.2Model Accessibility=Proprietary2026.04 | 84 | 89 | 86 | 81 | |
| GPT-5-MiniModel Accessibility=Proprietary2026.04 | 75 | 87 | 80 | 94 | |
| Llama-4-17BModel Accessibility=Open-weight2026.04 | 73 | 74 | 73 | 75 | |
| Nemotron-Nano-9BModel Accessibility=Open-weight2026.04 | 45 | 61 | 52 | 67 |