Multimodal Reasoning on AI2D
0.857ScoreLlava-CoT
Evaluation results
| Method | Links | |
|---|---|---|
| Llava-CoTModel Source=Open-Source2025.01 | 0.857 | |
| GPT-4o-0806Model Source=Close-Source2025.01 | 0.847 | |
| Ovis1.5-Gemma2-9BModel Source=Open-Source2025.01 | 0.845 | |
| LLaVA-OVScale=1.5-8B2026.06 | 0.8416 | |
| InternVL2-8BModel Source=Open-Source2025.01 | 0.836 | |
| Qwen2.5-VLScale=7B2026.06 | 0.8303 | |
| MiniCPM-V2.6-8BModel Source=Open-Source2025.01 | 0.821 | |
| LlamaV-o1Model Source=Our Models2025.01 | 0.8124 | |
| Claude3.5-Sonnet-0620Model Source=Close-Source2025.01 | 0.802 | |
| Gemini-1.5-ProModel Source=Close-Source2025.01 | 0.791 | |
| Mulberry-7BModel Source=Open-Source2025.01 | 0.7895 | |
| GPT-4o-mini-0718Model Source=Close-Source2025.01 | 0.778 | |
| VILA-1.5-40BModel Source=Open-Source2025.01 | 0.778 | |
| MOSS-Video-PreviewSFT Mode=real-time2026.06 | 0.7733 | |
| Llama-3.2-11B-Vision-InstModel Source=Our Models, Note=baseline2025.01 | 0.773 | |
| Llama-3.2 11B-VisionCheckpoint=base2026.06 | 0.7646 | |
| MOSS-Video-PreviewSFT Mode=offline2026.06 | 0.7506 | |
| Llama-3.2-90B-Vision-InstModel Source=Open-Source2025.01 | 0.695 |