NumberQA on ChartBench (test)
73.25Relaxed AccuracyChartAgent (Qwen3vl-235b)
Evaluation Results
| Method | Links | |
|---|---|---|
| ChartAgent (Qwen3vl-235b)Model Category=General MLLMs, Base Model=Qwen3VL-235b2025.12 | 73.25 | |
| Qwen3VL-235bModel Category=General MLLMs2025.12 | 71.48 | |
| ChartAgent (Qwen3vl-30b)Model Category=General MLLMs, Base Model=Qwen3VL-30b2025.12 | 68.46 | |
| Claude4Model Category=General MLLMs2025.12 | 63.86 | |
| Qwen3VL-30bModel Category=General MLLMs2025.12 | 63.81 | |
| Gemini-2.5 FlashModel Category=General MLLMs2025.12 | 63.26 | |
| Qwen2.5VL-72bModel Category=General MLLMs2025.12 | 63.11 | |
| ChartMoEModel Category=Chart Models2025.12 | 51.67 | |
| TinyChartModel Category=Chart Models2025.12 | 47.86 | |
| GPT-4oModel Category=General MLLMs2025.12 | 45.33 | |
| ChartGemmaModel Category=Chart Models2025.12 | 38.46 | |
| ChartVLMModel Category=Chart Models2025.12 | 32.19 | |
| ChartInstructModel Category=Chart Models2025.12 | 31.75 | |
| UniChartModel Category=Chart Models2025.12 | 27.44 | |
| MatChaModel Category=Chart Models2025.12 | 25.86 | |
| ChartLlamaModel Category=Chart Models2025.12 | 21.31 |