Chart VQA on ChartQAPro
55.8AccuracyClaude 3.5 Sonnet
Evaluation results
| Method | Links | |
|---|---|---|
| Claude 3.5 SonnetZero-shot=true2026.02 | 55.8 | |
| Gemini 2.0 FlashZero-shot=true2026.02 | 53.7 | |
| Qwen2.5-VL-32B + AT-RL (Ours)Zero-shot=true2026.02 | 48.4 | |
| Qwen2.5-VL-72B InstructZero-shot=true2026.02 | 47.2 | |
| OpenAI GPT-4oZero-shot=true2026.02 | 46.9 | |
| Qwen2.5-VL-32B + VPPOZero-shot=true2026.02 | 46.6 | |
| Socratic-Geo2026.02 | 46.3 | |
| Qwen2.5-VL-32B InstructZero-shot=true2026.02 | 43.4 | |
| Base Model2026.02 | 41.3 |