Multi-modal Visualization Generation on MultiVis-Bench Image-Referenced Generation
80.33LL: TypeMultiVis-Agent
Evaluation results
| Method | Links | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MultiVis-AgentModel=gemini2026.01 | 80.33 | 45.19 | 38.91 | 94.14 | 76.99 | 79.08 | 69.11 | 89.85 | 90.96 | 75.84 | 74.06 | 80.75 | 95.06 | 82.16 | 75.63 | |
| MultiVis-AgentModel=gpt2026.01 | 71.97 | 38.08 | 20.5 | 90.38 | 69.87 | 76.99 | 61.3 | 80.33 | 81.09 | 68.93 | 61.51 | 63.6 | 91.05 | 71.84 | 66.57 | |
| LLM WorkflowModel=gemini2026.01 | 64.85 | 34.73 | 22.18 | 85.77 | 46.86 | 68.62 | 53.84 | 81.38 | 78.24 | 66.42 | 63.08 | 68.1 | 83.01 | 71.75 | 62.79 | |
| LLM WorkflowModel=gpt2026.01 | 61.92 | 33.05 | 10.46 | 68.62 | 24.69 | 63.18 | 43.65 | 67.47 | 66.44 | 57.64 | 50.73 | 52.93 | 72.01 | 59.35 | 51.5 | |
| Instructing LLMModel=gemini2026.01 | 59.83 | 39.75 | 20.5 | 81.59 | 38.08 | 62.76 | 50.42 | 71.13 | 69.41 | 61.09 | 57.32 | 62.55 | 77.2 | 64.66 | 57.54 | |
| Instructing LLMModel=gpt2026.01 | 58.16 | 40.17 | 16.32 | 74.48 | 27.62 | 57.74 | 45.75 | 66.32 | 66.95 | 58.37 | 51.78 | 55.96 | 73.1 | 60.41 | 53.08 | |
| nvAgentModel=gpt2026.01 | 30.54 | 15.06 | 2.51 | 65.27 | 64.85 | 51.88 | 38.35 | 30.33 | 30.46 | 27.82 | 18.41 | 32.11 | 50.25 | 29.68 | 34.02 | |
| nvAgentModel=gemini2026.01 | 26.78 | 20.08 | 2.51 | 57.74 | 56.07 | 46.03 | 34.87 | 33.26 | 31.59 | 29.92 | 22.49 | 33.05 | 49.33 | 31.79 | 33.33 |