Diagram Creation on CanvasBench (test)
4.201CCFGemini-3-Pro
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Gemini-3-ProCategory=LATEXTikz2026.02 | 4.201 | 4.129 | 4.086 | 3.368 | 3.67 | 3.641 | 3.124 | 3.871 | 3.416 | 2.541 | |
| GPT-5.2Category=LATEXTikz2026.02 | 4.096 | 3.933 | 4.067 | 2.833 | 3.257 | 2.933 | 2.543 | 3.295 | 3.171 | 2.505 | |
| EvoDiagramBackbone=GPT-5.2, Protocol=w/o Multiagent2026.02 | 3.97 | 4.02 | 4.1 | 2.428 | 3.025 | 2.289 | 1.891 | 2.662 | 3.189 | 2.786 | |
| GPT-5.2Category=Prompting2026.02 | 3.914 | 3.922 | 4.025 | 2.312 | 2.945 | 2.184 | 1.782 | 2.584 | 3.146 | 2.712 | |
| EvoDiagramBackbone=GPT-5.2, Protocol=Standard2026.02 | 3.904 | 3.847 | 4.158 | 2.929 | 3.224 | 2.405 | 2.705 | 2.843 | 3.419 | 2.962 | |
| EvoDiagramBackbone=Gemini-3-Pro, Protocol=w/o Multiagent2026.02 | 3.888 | 3.995 | 3.961 | 2.952 | 3.275 | 3.14 | 2.435 | 3.387 | 3.135 | 2.705 | |
| EvoDiagramBackbone=GPT-5.2, Protocol=w/o Memory2026.02 | 3.882 | 3.825 | 4.135 | 2.894 | 3.186 | 2.378 | 2.678 | 2.821 | 3.397 | 2.94 | |
| EvoDiagramBackbone=Gemini-3-Pro, Protocol=Standard2026.02 | 3.853 | 3.863 | 3.887 | 3.177 | 3.26 | 3.294 | 3.01 | 3.52 | 3.196 | 2.657 | |
| EvoDiagramBackbone=Gemini-3-Pro, Protocol=w/o Memory2026.02 | 3.832 | 3.841 | 3.865 | 3.142 | 3.218 | 3.261 | 2.978 | 3.495 | 3.174 | 2.635 | |
| Gemini-3-ProCategory=Prompting2026.02 | 3.765 | 3.904 | 3.821 | 2.842 | 3.124 | 2.986 | 2.312 | 3.214 | 3.042 | 2.564 | |
| EvoDiagramBackbone=Qwen3-VL-Plus, Protocol=Standard2026.02 | 3.654 | 3.612 | 3.748 | 2.852 | 3.116 | 2.873 | 2.695 | 3.042 | 3.018 | 2.406 | |
| EvoDiagramBackbone=Qwen3-VL-Plus, Protocol=w/o Memory2026.02 | 3.632 | 3.591 | 3.725 | 2.818 | 3.082 | 2.839 | 2.662 | 3.018 | 2.996 | 2.384 | |
| Kimi-K2-0905Category=LATEXTikz2026.02 | 3.603 | 3.52 | 3.564 | 2.461 | 2.835 | 2.738 | 2.112 | 2.966 | 2.607 | 1.942 | |
| GLM-4.7Category=LATEXTikz2026.02 | 3.569 | 3.512 | 3.507 | 2.762 | 3.224 | 2.876 | 2.491 | 3.276 | 2.829 | 2.191 | |
| Qwen3-VL-PlusCategory=LATEXTikz2026.02 | 3.492 | 3.3 | 3.437 | 2.206 | 2.688 | 2.427 | 1.744 | 2.598 | 2.362 | 1.653 | |
| EvoDiagramBackbone=Qwen3-VL-Plus, Protocol=w/o Multiagent2026.02 | 3.412 | 3.395 | 3.48 | 2.365 | 2.89 | 2.412 | 1.884 | 2.585 | 2.392 | 1.854 | |
| GLM-4.7Category=Prompting2026.02 | 3.392 | 3.424 | 3.485 | 2.452 | 2.912 | 2.595 | 2.014 | 2.784 | 2.516 | 1.942 | |
| Kimi-K2-0905Category=Prompting2026.02 | 3.305 | 3.242 | 3.314 | 2.254 | 2.965 | 2.384 | 1.805 | 2.586 | 2.312 | 1.684 | |
| Qwen3-VL-PlusCategory=Prompting2026.02 | 3.284 | 3.256 | 3.354 | 2.186 | 2.721 | 2.214 | 1.705 | 2.392 | 2.224 | 1.712 | |
| GPT-4o-ImageCategory=Diffusion2026.02 | 2.2 | 3.11 | 2.71 | 3.99 | 4.354 | 4.153 | 3.923 | 4.301 | 2.751 | 3.043 | |
| NanoBananoCategory=Diffusion2026.02 | 1.986 | 2.871 | 2.657 | 3.767 | 3.838 | 3.533 | 3.671 | 3.761 | 2.665 | 2.99 | |
| Flux.2 flexCategory=Diffusion2026.02 | 1.871 | 2.782 | 2.643 | 3.836 | 3.868 | 3.614 | 3.704 | 3.78 | 2.712 | 2.922 |