Chart-to-code generation on Chart2NCode Chart2R (test)
13Error Rate (ER)ChartLlama-13B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ChartLlama-13BModel Category=Open-source2026.04 | 13 | 44.8 | 6.2 | |
| DeepSeek-VL2-3BModel Category=Open-source2026.04 | 23 | 56.9 | 16.4 | |
| ChartMoE-8BModel Category=Open-source2026.04 | 39.3 | 52.9 | 25.5 | |
| Phi-3.5-vision-4BModel Category=Open-source2026.04 | 47 | 52.5 | 33.5 | |
| Qwen3-VL-2BModel Category=Open-source2026.04 | 56.5 | 52.4 | 42 | |
| LLaVA-v1.6-7BModel Category=Open-source2026.04 | 58.5 | 50.9 | 37.9 | |
| InternVL3.5-2BModel Category=Open-source2026.04 | 61.8 | 53.4 | 44.9 | |
| InternVL3.5-4BModel Category=Open-source2026.04 | 66.8 | 56.4 | 51.5 | |
| InternVL3.5-8BModel Category=Open-source2026.04 | 67 | 67.6 | 48.2 | |
| Qwen3-VL-8BModel Category=Open-source2026.04 | 73.6 | 72.7 | 57.2 | |
| Qwen3-VL-4BModel Category=Open-source2026.04 | 75.4 | 66.4 | 60.9 | |
| GPT-4o-miniModel Category=Proprietary2026.04 | 89.5 | 75.4 | 70.3 | |
| GPT-5-miniModel Category=Proprietary2026.04 | 90.3 | 82.5 | 81.2 | |
| Claude-Haiku-3.5Model Category=Proprietary2026.04 | 93 | 76.2 | 73.9 | |
| Claude-Sonnet-4Model Category=Proprietary2026.04 | 93.9 | 82 | 83.1 | |
| GPT-4oModel Category=Proprietary2026.04 | 94.5 | 78.8 | 78.3 | |
| CharLuMA-1.3BModel Category=Open-source2026.04 | 94.5 | 78.9 | 73.3 | |
| CharLuMA-6.7BModel Category=Open-source2026.04 | 96.5 | 81.8 | 80.9 |