Chart Question Answering on ChartQAPRO
66.8Overall AccuracyQwen3.5-27B
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Qwen3.5-27BMode=REASONING, Architecture=Dense, # Total Params=27B, # Activated Params=27B2026.04 | 66.8 | — | — | — | — | — | — | |
| EXAONE 4.5 33BMode=REASONING, Architecture=Dense, # Total Params=33B, # Activated Params=33B2026.04 | 62.2 | — | — | — | — | — | — | |
| Qwen3-VL-32BMode=Thinking, Architecture=Dense, # Total Params=33B, # Activated Params=33B2026.04 | 61.4 | — | — | — | — | — | — | |
| Qwen3-VL-235B-A22BMode=Thinking, Architecture=MoE, # Total Params=236B, # Activated Params=23B2026.04 | 61.2 | — | — | — | — | — | — | |
| GPT-5 miniMode=REASONING: HIGH, Architecture=-, # Total Params=-, # Activated Params=-2026.04 | 60.9 | — | — | — | — | — | — | |
| Gemini-Flash-2.02026.03 | 46.85 | 43.43 | 60.28 | 40.25 | 67.62 | 24.47 | 47.15 | |
| Gemini-Flash-2.02026.03 | 46.85 | 43.43 | 60.28 | 40.25 | 67.62 | 24.47 | — | |
| Claude-3.5 Sonnet2026.03 | 43.58 | 38.84 | 51.4 | 44.53 | 55.6 | 45.48 | 46.57 | |
| Qwen2.5-VL-7B-InstructOptimization Strategy=PRPO2026.03 | 42.95 | 36.24 | 50.47 | 49.63 | 53.28 | 53.69 | 47.69 | |
| Qwen2.5-VL-7B-InstructOptimization Strategy=ChartReasoner-GRPO2026.03 | 39.97 | — | — | — | — | — | — | |
| Qwen2.5-VL-7B-InstructOptimization Strategy=ChartReasoner-SFT2026.03 | 37.94 | — | — | — | — | — | — | |
| GPT-4o2026.03 | 37.67 | 35.76 | 46.72 | 34.75 | 45.49 | 28.91 | 38.33 | |
| GPT-4o2026.03 | 37.67 | 35.76 | 46.72 | 34.75 | 45.49 | 28.91 | — | |
| Qwen2.5-VL-7B-InstructOptimization Strategy=Baseline2026.03 | 36.31 | 27.49 | 37.85 | 55.22 | 46.72 | 44.4 | 41.33 | |
| InternVL2.5-8B2026.03 | 35.67 | 35.21 | 25.7 | 32.26 | 53.27 | 29.61 | 35.29 | |
| VisDoT_InternVLParam=4B2026.03 | 34.54 | 32.75 | 47.2 | 15.38 | 53.28 | 40.82 | — | |
| InternVLParam=4B2026.03 | 32.11 | 25.81 | 10.98 | 15.85 | 47.13 | 25.07 | — | |
| ChartGemmaParam=3B2026.03 | 24.86 | 6.86 | 0 | 16 | 1.22 | 6.53 | — | |
| VisDoT_InternVLParam=2B2026.03 | 24.35 | 23.4 | 29.91 | 12.03 | 37.3 | 29.59 | — | |
| LLaVA-Next-7B2026.03 | 21.97 | 15.35 | 35.98 | 21.09 | 41.8 | 17.79 | 25.83 | |
| InternVLParam=2B2026.03 | 17.81 | 13.86 | 10.74 | 14.02 | 45.9 | 18.92 | — | |
| TinyChart-3B2026.03 | 13.25 | 8.52 | 7 | 17.46 | 33.19 | 16.06 | 16.45 | |
| Llama-3.2-11B-Vision2026.03 | 11.09 | 12.34 | 2.33 | 0.19 | 27.18 | 10.93 | 10.68 | |
| ChartGemma-3B2026.03 | 6.84 | 6.86 | 0 | 16 | 1.22 | 6.53 | 6.12 | |
| ChartInstruct-LLama2-7B2026.03 | 4.88 | 7.09 | 0 | 3.77 | 0 | 6.91 | 3.55 |