High-Resolution Fine-Grained Visual Perception on HR-Bench-8K
81.5Overall ScoreGemini-2.5-Pro
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Gemini-2.5-ProModel Category=Proprietary Model2026.06 | 81.5 | 83 | 80 | |
| Qwen3-VL-235B-A22BModel Category=Open-Source Models, Backbone=235B-A22B2026.06 | 80.4 | 83 | 77.3 | |
| Imagine-OPD-8BModel Category=Our Models, Backbone=8B2026.06 | 79.5 | 88.3 | 70.8 | |
| Imagine-OPD-4BModel Category=Our Models, Backbone=4B2026.06 | 78.6 | 88 | 69.3 | |
| Kimi-K2.5 (1T)Model Category=Open-Source Models, Backbone=1T2026.06 | 75.1 | 82.8 | 67.5 | |
| Qwen3-VL-8BModel Category=Open-Source Models, Backbone=8B2026.06 | 74.6 | 82.3 | 66.8 | |
| Qwen3-VL-30B-ThinkingModel Category=Open-Source Models, Backbone=30B2026.06 | 74.2 | 80.3 | 68.3 | |
| Gemini-2.5-FlashModel Category=Proprietary Model2026.06 | 73.7 | 75.8 | 71.8 | |
| Gemini-3-FlashModel Category=Proprietary Model2026.06 | 73.7 | 75.8 | 71.8 | |
| TreeVGR-7BModel Category=Reasoning with Tool Methods, Backbone=7B2026.06 | 73.1 | 86.5 | 59.6 | |
| Qwen3-VL-4BModel Category=Open-Source Models, Backbone=4B2026.06 | 72.9 | 79.3 | 66.3 | |
| DeepEyesModel Category=Reasoning with Tool Methods2026.06 | 72.6 | 86.8 | 58.5 | |
| ThymeModel Category=Reasoning with Tool Methods2026.06 | 72 | 72 | 57.5 | |
| Qwen2.5-VL-32BModel Category=Open-Source Models, Backbone=32B2026.06 | 70.4 | 82.3 | 58.5 | |
| InternVL3-8BModel Category=Open-Source Models, Backbone=8B2026.06 | 69.3 | 78.8 | 59.8 | |
| Pixel-ReasonerModel Category=Reasoning with Tool Methods2026.06 | 66.9 | 80 | 54.3 | |
| Qwen2.5-VL-7BModel Category=Open-Source Models, Backbone=7B2026.06 | 63.8 | 73.8 | 53.8 | |
| GPT-4oModel Category=Proprietary Model2026.06 | 59.6 | 60.8 | 58.5 |