Chart Reasoning on CharXiv
74.2AccuracySocratic-Geo
Evaluation results
| Method | Links | |
|---|---|---|
| Socratic-Geo2026.02 | 74.2 | |
| Base Model2026.02 | 66.6 | |
| Qwen2.5-VL-72BModel Category=Open-Source MLLMs2025.12 | 49.7 | |
| Qwen2.5-VL-32BModel Category=Open-Source MLLMs2025.12 | 47.6 | |
| GPT-4oModel Category=Closed-Source MLLMs2025.12 | 47.1 | |
| InternVL3-78BModel Category=Open-Source MLLMs2025.12 | 46 | |
| Thyme-VL-7BModel Category=Open-Source MLLMs with Tools, reproduced=true2025.12 | 44.2 | |
| CodeDance-7BModel Category=Open-Source MLLMs with Tools2025.12 | 44.1 | |
| Qwen2.5-VL-7BModel Category=Open-Source MLLMs2025.12 | 42.1 | |
| InternVL3-8BModel Category=Open-Source MLLMs2025.12 | 38.3 | |
| InternVL2.5-8BModel Category=Open-Source MLLMs2025.12 | 37.2 | |
| Llava-OneVision-7BModel Category=Open-Source MLLMs2025.12 | 27.1 |