Visual Reasoning on VSPO
71.08AccuracyAdaReasoner 7B
Evaluation results
| Method | Links | |
|---|---|---|
| AdaReasoner 7BModel Category=Tool-Planning Models, Backbone=7B2026.01 | 71.08 | |
| GPT 5 + ToolsModel Category=Tool-Planning Models, Tool Usage=true2026.01 | 52.75 | |
| Claude 4 sonnetModel Category=Closed-Source Models2026.01 | 51.56 | |
| Gemini 2.5 flashModel Category=Closed-Source Models2026.01 | 40.12 | |
| Qwen 2.5 VL 72B + ToolsModel Category=Tool-Planning Models, Tool Usage=true, Backbone=72B2026.01 | 39.64 | |
| GPT 5Model Category=Closed-Source Models2026.01 | 34.25 | |
| Qwen 2.5 VL 72BModel Category=Open-Source Models, Backbone=72B2026.01 | 33.41 | |
| Qwen 2.5 VL 7B + ToolsModel Category=Tool-Planning Models, Tool Usage=true, Backbone=7B2026.01 | 28.98 | |
| Qwen 2.5 VL 32BModel Category=Open-Source Models, Backbone=32B2026.01 | 28.56 | |
| InternVL3 78BModel Category=Open-Source Models, Backbone=78B2026.01 | 28.14 | |
| Qwen 2.5 VL 3BModel Category=Open-Source Models, Backbone=3B2026.01 | 26.53 | |
| Qwen 2.5 VL 7BModel Category=Open-Source Models, Backbone=7B2026.01 | 25.39 | |
| PixelReasonerModel Category=Tool-Planning Models, Unified Evaluation Framework=true2026.01 | 21.38 | |
| DeepEyesModel Category=Tool-Planning Models, Unified Evaluation Framework=true2026.01 | 10.18 |