Multi-Image Question Answering on Molmo QA Benchmarks Multi-Image 19
81.9Average Score (Multi-Image)Gemini 3 Pro
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini 3 ProAccess=API call only2026.03 | 81.9 | |
| Gemini 2.5 ProAccess=API call only2026.03 | 72.4 | |
| GPT-5Access=API call only2026.03 | 72.1 | |
| Gemini 2.5 FlashAccess=API call only2026.03 | 68.3 | |
| GPT-5 miniAccess=API call only2026.03 | 68.2 | |
| GLM-4.1V-9BAccess=Open weights2026.03 | 67.4 | |
| Claude Sonnet 4.5Access=API call only2026.03 | 59.5 | |
| Molmo2-4BAccess=Fully open2026.03 | 57.8 | |
| Qwen3-VL-4BAccess=Open weights2026.03 | 57.6 | |
| Molmo2-8BAccess=Fully open2026.03 | 56.4 | |
| Qwen3-VL-8BAccess=Open weights2026.03 | 56.3 | |
| MolmoPoint-8BAccess=MolmoPoint2026.03 | 56 | |
| InternVL3.5-8BAccess=Open weights2026.03 | 54.9 | |
| InternVL3.5-4BAccess=Open weights2026.03 | 53.5 | |
| Molmo2-O-7BAccess=Fully open2026.03 | 53.5 | |
| Keye-VL-1.5-8BAccess=Open weights2026.03 | 52.1 | |
| Eagle2.5-8BAccess=Open weights2026.03 | 52 | |
| MiniCPM-V-4.5-8BAccess=Open weights2026.03 | 47.3 | |
| PLM-3BAccess=Fully open2026.03 | 40.6 | |
| PLM-8BAccess=Fully open2026.03 | 35.7 |