ScreenSpot-Pro
专业软件高分辨率截图上的界面定位，覆盖开发、创意、CAD、科学、办公和操作系统 6 类软件，分 Text 与 Icon 目标统计。
收录成绩9999 个模型或系统
当前第一85.1GroundCascade (Indeed-UI-32B + KV-Ground-8B)
指标定位准确率定位正确的样本占比（%），榜单注明按样本微平均
最近同步10-16 16:571 个数据源 · 界面定位
排名表
显示 99 / 99
| 总体 | 开发 | 创意 | CAD | 科学 | 办公 | 操作系统 | |||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 模型 | 机构 | 参数 | 开源 | 评测设置 | 来源 | 备注 | |||||||||||||||||
| 1 | GroundCascade (Indeed-UI-32B + KV-Ground-8B) | — | — | — | 85.1 | 92.1 | 73.8 | 93.5 | 80.7 | 85.3 | 74.1 | 92.4 | 70.3 | 95.1 | 72.7 | 96.6 | 79.3 | 90.7 | 62.9 | — | 官方榜单 | 2026-10-16 | Training-free geometric consistency cascade; Indeed-UI-32B primary/verifier + conditional KV-Ground-8B helper; fixed 0.5 zoom; 1% self-consistency gate; 4% cross-model agreement gate; greedy decoding; 2.574 average model calls/example. |
| 2 | GroundCascade (Indeed-UI-32B + Indeed-UI-8B) | — | — | — | 85.0 | 92.4 | 73.0 | 93.5 | 78.6 | 85.9 | 74.1 | 92.9 | 70.3 | 95.1 | 70.9 | 96.6 | 79.3 | 91.6 | 62.9 | — | 官方榜单 | 2026-10-16 | Training-free geometric consistency cascade; Indeed-UI-32B primary/verifier + conditional Indeed-UI-8B helper; fixed 0.5 zoom; 1% self-consistency gate; 4% cross-model agreement gate; greedy decoding; 2.572 average model calls/example. |
| 3 | Indeed-UI-32B-zoomin | — | — | — | 82.7 | 90.4 | 70.4 | 90.9 | 75.9 | 86.4 | 71.3 | 89.8 | 65.6 | 92.4 | 69.1 | 97.2 | 79.3 | 84.1 | 59.5 | — | 官方榜单 | 2026-10-16 | |
| 4 | Indeed-UI-8B-zoom-in | — | — | — | 81.0 | 89.9 | 66.6 | 92.2 | 77.2 | 86.4 | 62.2 | 89.8 | 62.5 | 91.7 | 62.7 | 94.3 | 73.6 | 83.2 | 59.5 | — | 官方榜单 | 2026-10-16 | |
| 5 | KV-Ground-8B + Qwen3.5-27B Consistency Router | — | — | — | 80.9 | 90.4 | 65.6 | 93.5 | 73.1 | 85.3 | 62.2 | 87.8 | 51.6 | 93.8 | 70.9 | 94.9 | 75.5 | 87.8 | 56.2 | — | 官方榜单 | 2026-10-16 | Heterogeneous ensemble of KV-Ground-8B-BaseGuiOwl1.5-0315 and Qwen3.5-27B-AWQ-4bit. Both models run independent 2-step zoom-in pipelines (crop_ratio=0.5). Per sample, the model with lower zoom consistency (step-2 prediction closer to crop center = higher confidence) is selected. KV-Ground selected … |
| 6 | KV-Ground-GuiOwl1.5-0315-8B-ZoomIn | — | — | — | 80.5 | 90.0 | 65.2 | 90.9 | 71.7 | 86.4 | 60.8 | 90.4 | 60.9 | 93.1 | 67.3 | 93.8 | 73.6 | 84.1 | 57.3 | — | 官方榜单 | 2026-10-16 | |
| 7 | Holo2-235B-A22B (Agentic) | — | — | — | 78.5 | 78.8 | 78.0 | 73.4 | 75.9 | 70.2 | 74.8 | 80.7 | 79.7 | 80.6 | 79.1 | 85.9 | 83.0 | 85.0 | 80.9 | — | 官方榜单 | 2026-10-16 | Holo2-235B-A22B (Agentic) |
| 8 | MAI-UI-32B (MVP) | — | — | — | 77.5 | 88.3 | 60.0 | 92.9 | 64.8 | 83.5 | 50.0 | 83.8 | 54.7 | 92.4 | 54.5 | 92.7 | 86.8 | 86.0 | 62.9 | — | 官方榜单 | 2026-10-16 | MVP on attention layer 48; max inference process is 2. |
| 9 | AdaZoom-GUI-4B (cond. zoom-in, refine w/ Qwen3.5-397B) | — | — | — | 76.8 | 90.1 | 55.3 | 91.6 | 57.9 | 88.9 | 49.6 | 83.3 | 39.1 | 91.0 | 51.8 | 97.2 | 81.1 | 89.7 | 60.7 | — | 官方榜单 | 2026-10-16 | Conditional zoom-in and refine with Qwen3.5-397B-A17B |
| 10 | KV-Ground-GuiOwl1.5-4B-0228-ZoomIn | — | — | — | 76.4 | 85.6 | 61.6 | 88.3 | 57.9 | 82.3 | 58.7 | 81.7 | 53.1 | 88.9 | 61.8 | 92.7 | 81.1 | 78.5 | 66.3 | — | 官方榜单 | 2026-10-16 | |
| 11 | Holo2-30B-A3B (Agentic) | — | — | — | 75.2 | 76.5 | 73.2 | 70.1 | 72.4 | 68.2 | 67.1 | 74.1 | 68.8 | 81.9 | 80.9 | 84.8 | 77.4 | 84.1 | 75.3 | — | 官方榜单 | 2026-10-16 | Holo2-30B-A3B (Agentic) |
| 12 | MVP_Qwen3VL-32B | — | — | — | 74.2 | 88.0 | 51.7 | 90.9 | 51.0 | 87.4 | 44.8 | 83.3 | 48.4 | 89.6 | 45.5 | 91.5 | 75.5 | 86.0 | 59.5 | — | 官方榜单 | 2026-10-16 | MVP crops different regions from original screenshot to get different prediction and then cluster them through k-means. The final result is decided by the centroid coordinate of the largest cluster. |
| 13 | MAI-UI-32B (Zoom In) | — | — | — | 73.5 | 85.4 | 54.3 | 84.4 | 57.9 | 87.9 | 46.1 | 79.2 | 53.1 | 91.7 | 54.5 | 88.1 | 79.3 | 80.4 | 47.2 | — | 官方榜单 | 2026-10-16 | |
| 14 | Indeed-UI-8B | — | — | — | 73.4 | 83.6 | 56.8 | 86.4 | 69.7 | 82.8 | 48.3 | 74.6 | 51.6 | 86.8 | 52.7 | 93.8 | 64.2 | 76.6 | 53.9 | — | 官方榜单 | 2026-10-16 | |
| 15 | Indeed-UI-32B | — | — | — | 73.3 | 85.3 | 54.0 | 90.3 | 66.2 | 81.3 | 46.9 | 78.2 | 51.6 | 91.0 | 50.9 | 92.7 | 66.0 | 78.5 | 43.8 | — | 官方榜单 | 2026-10-16 | |
| 16 | KV-Ground-GuiOwl1.5-0315-8B | — | — | — | 73.2 | 83.5 | 56.5 | 86.4 | 69.7 | 81.8 | 49.0 | 75.6 | 54.7 | 86.8 | 49.1 | 93.8 | 62.3 | 75.7 | 53.9 | — | 官方榜单 | 2026-10-16 | |
| 17 | Duvo Eye-1 (Holo-3.1-35B-A3B + LoRA) | — | — | — | 72.9 | 81.4 | 59.3 | 84.4 | 67.6 | 79.3 | 49.0 | 71.6 | 45.3 | 82.6 | 54.5 | 89.3 | 81.1 | 84.1 | 65.2 | — | 模型方自报 | 2026-10-16 | Holo-3.1-35B-A3B with a LoRA adapter (rank 64, alpha 128, 1 epoch) trained on duvoai/SynthUI (14.9k rows). Single-shot grounding, temperature 0, thinking disabled, guided-JSON {"x","y"} in [0,1000] scaled to absolute pixels. Self-reported; per-sample predictions published for rescoring at https://h… |
| 18 | HuzzleWorld-3B (Zoom In) | — | — | — | 72.4 | 79.2 | 61.4 | 82.5 | 62.8 | 78.8 | 56.6 | 65.5 | 51.6 | 84.7 | 61.8 | 87.6 | 81.1 | 79.4 | 61.8 | — | 官方榜单 | 2026-10-16 | Proprietary |
| 19 | MAI-UI-8B (Zoom In) | — | — | — | 71.9 | 82.1 | 55.5 | 78.6 | 61.4 | 78.8 | 49.0 | 83.3 | 45.3 | 87.5 | 52.7 | 87.0 | 81.1 | 75.7 | 51.7 | — | 官方榜单 | 2026-10-16 | |
| 20 | Holo2-8B (Agentic) | — | — | — | 71.4 | 71.0 | 72.0 | 69.5 | 74.5 | 63.6 | 64.3 | 68.0 | 65.6 | 76.4 | 79.1 | 75.7 | 73.6 | 77.6 | 75.3 | — | 官方榜单 | 2026-10-16 | Holo2-8B (Agentic) |
| 21 | HuzzleWorld-2B (Zoom In) | — | — | — | 70.8 | 80.3 | 55.5 | 83.1 | 65.5 | 74.8 | 49.6 | 76.1 | 51.6 | 82.6 | 50.0 | 88.1 | 64.2 | 78.5 | 52.8 | — | 官方榜单 | 2026-10-16 | Proprietary |
| 22 | AdaZoom-GUI-4B (conditional zoom-in) | — | — | — | 70.6 | 85.4 | 46.7 | 87.7 | 47.6 | 83.3 | 38.5 | 77.7 | 32.8 | 88.9 | 49.1 | 94.3 | 66.0 | 80.4 | 53.9 | — | 官方榜单 | 2026-10-16 | |
| 22 | Holo2-235B-A22B | — | — | — | 70.6 | 70.1 | 71.4 | 70.8 | 72.4 | 59.1 | 64.3 | 67.5 | 67.2 | 69.4 | 75.5 | 81.4 | 71.7 | 76.6 | 78.7 | — | 官方榜单 | 2026-10-16 | Holo2-235B-A22B |
| 24 | KV-Ground-Qwen3VL-4B-ZoomIn | — | — | — | 70.3 | 84.3 | 47.5 | 87.7 | 44.1 | 80.3 | 41.3 | 75.6 | 39.1 | 90.3 | 50.0 | 92.7 | 67.9 | 81.3 | 53.9 | — | 官方榜单 | 2026-10-16 | |
| 25 | UI-Venus-1-5-30B-A3B | — | — | — | 69.6 | 81.2 | 51.0 | 87.7 | 57.9 | 75.8 | 41.3 | 70.6 | 40.6 | 84.0 | 47.3 | 89.8 | 69.8 | 83.2 | 56.2 | — | 官方榜单 | 2026-10-16 | |
| 26 | Holo2-4B (Agentic) | — | — | — | 68.6 | 69.4 | 67.2 | 65.6 | 71.7 | 59.6 | 58.0 | 64.5 | 57.8 | 79.2 | 72.7 | 76.3 | 73.6 | 77.6 | 70.8 | — | 官方榜单 | 2026-10-16 | Holo2-4B (Agentic) |
| 27 | UI-Venus-1-5-8B | — | — | — | 68.4 | 82.4 | 45.9 | 85.7 | 54.5 | 75.3 | 32.9 | 75.1 | 31.3 | 86.1 | 44.5 | 92.7 | 66.0 | 82.2 | 52.8 | — | 官方榜单 | 2026-10-16 | |
| 28 | MAI-UI-32B | — | — | — | 67.9 | 83.0 | 43.4 | 86.4 | 40.7 | 82.8 | 37.8 | 70.0 | 45.3 | 91.7 | 46.4 | 90.4 | 71.7 | 78.5 | 34.8 | — | 官方榜单 | 2026-10-16 | |
| 29 | KV-Ground-GuiOwl1.5-0228-4B | — | — | — | 67.0 | 78.2 | 49.0 | 84.4 | 49.7 | 75.3 | 46.1 | 59.4 | 40.6 | 88.9 | 48.2 | 90.4 | 58.5 | 74.8 | 53.9 | — | 官方榜单 | 2026-10-16 | |
| 30 | HuzzleWorld-3B | — | — | — | 66.4 | 74.0 | 54.1 | 81.2 | 55.2 | 75.8 | 51.0 | 50.8 | 42.2 | 81.9 | 53.6 | 85.3 | 71.7 | 73.8 | 56.2 | — | 官方榜单 | 2026-10-16 | Proprietary |
| 31 | Holo2-30B-A3B | — | — | — | 66.1 | 80.3 | 43.0 | 87.7 | 51.0 | 77.8 | 37.8 | 62.4 | 31.3 | 86.1 | 36.4 | 90.4 | 52.8 | 83.2 | 49.4 | — | 官方榜单 | 2026-10-16 | |
| 32 | MAI-UI-8B | — | — | — | 65.7 | 79.9 | 42.5 | 85.1 | 51.0 | 76.3 | 29.4 | 73.1 | 42.2 | 77.1 | 36.4 | 89.3 | 56.6 | 80.4 | 49.4 | — | 官方榜单 | 2026-10-16 | |
| 33 | MVP_Qwen3VL-8B | — | — | — | 65.0 | 80.5 | 39.9 | 83.8 | 37.2 | 76.1 | 36.4 | 71.6 | 34.4 | 85.4 | 40.0 | 90.4 | 56.6 | 77.6 | 43.8 | — | 官方榜单 | 2026-10-16 | MVP crops different regions from original screenshot to get different prediction and then cluster them through k-means. The final result is decided by the centroid coordinate of the largest cluster. |
| 34 | GTA1-32B | — | — | — | 63.6 | 78.9 | 38.9 | 83.1 | 37.9 | 72.2 | 25.9 | 70.0 | 31.3 | 84.7 | 39.1 | 89.3 | 64.2 | 76.6 | 51.7 | — | 官方榜单 | 2026-10-16 | |
| 35 | HuzzleWorld-1B (Zoom In) | — | — | — | 63.5 | 74.6 | 45.5 | 76.6 | 51.0 | 74.8 | 41.3 | 66.0 | 37.5 | 82.6 | 39.1 | 79.7 | 60.4 | 68.2 | 48.3 | — | 官方榜单 | 2026-10-16 | Proprietary |
| 36 | Holo1.5-72B | — | — | — | 63.3 | 76.4 | 42.0 | 85.1 | 40.7 | 80.8 | 37.1 | 55.3 | 39.1 | 81.3 | 41.8 | 85.9 | 58.5 | 72.0 | 44.9 | — | 官方榜单 | 2026-10-16 | |
| 37 | KV-Ground-Qwen3VL-4B | — | — | — | 63.2 | 78.3 | 38.7 | 81.2 | 43.5 | 74.2 | 32.9 | 62.9 | 28.1 | 86.8 | 35.5 | 89.8 | 45.3 | 79.4 | 48.3 | — | 官方榜单 | 2026-10-16 | |
| 38 | MAI-UI-2B (Zoom In) | — | — | — | 62.8 | 76.5 | 40.7 | 75.3 | 42.8 | 74.8 | 30.1 | 69.5 | 34.4 | 84.0 | 42.7 | 85.3 | 56.6 | 69.2 | 47.2 | — | 官方榜单 | 2026-10-16 | |
| 39 | UI-Venus-72B | — | — | — | 61.9 | 77.4 | 36.8 | 84.4 | 33.1 | 73.2 | 30.8 | 66.5 | 29.7 | 84.7 | 42.7 | 83.0 | 60.4 | 75.7 | 36.0 | — | 官方榜单 | 2026-10-16 | |
| 40 | HuzzleWorld-2B | — | — | — | 61.7 | 70.8 | 47.0 | 74.0 | 53.8 | 69.7 | 40.6 | 56.4 | 37.5 | 75.7 | 42.7 | 81.9 | 54.7 | 70.1 | 53.9 | — | 官方榜单 | 2026-10-16 | Proprietary |
| 41 | AdaZoom-GUI-4B | — | — | — | 61.6 | 78.3 | 34.6 | 81.2 | 40.0 | 74.2 | 23.8 | 61.4 | 21.9 | 87.5 | 39.1 | 88.7 | 50.9 | 83.2 | 37.1 | — | 官方榜单 | 2026-10-16 | |
| 41 | UI-TARS-1.5 | — | — | — | 61.6 | 76.3 | 37.9 | 84.4 | 42.1 | 71.7 | 21.0 | 65.5 | 35.9 | 89.6 | 42.7 | 83.6 | 66.0 | 62.6 | 37.1 | — | 官方榜单 | 2026-10-16 | |
| 43 | Seed-1.5-VL | — | — | — | 60.9 | 78.2 | 33.0 | 76.6 | 29.7 | 79.8 | 30.8 | 68.5 | 29.7 | 84.0 | 31.8 | 84.2 | 43.4 | 77.6 | 39.3 | — | 官方榜单 | 2026-10-16 | |
| 44 | GUI-ARP-7B | — | — | — | 60.8 | 76.5 | 35.4 | 77.9 | 39.3 | 74.6 | 27.3 | 74.6 | 21.9 | 79.2 | 40.0 | 85.9 | 50.9 | 69.2 | 37.1 | — | 官方榜单 | 2026-10-16 | |
| 45 | GUI-AIMA-3B | — | — | — | 59.6 | 73.1 | 37.8 | 72.7 | 36.5 | 66.7 | 23.8 | 65.0 | 34.4 | 81.3 | 41.8 | 85.3 | 60.4 | 69.2 | 46.1 | — | 官方榜单 | 2026-10-16 | |
| 46 | Holo2-8B | — | — | — | 58.9 | 71.7 | 38.3 | 68.2 | 42.1 | 68.7 | 29.4 | 57.9 | 35.9 | 79.9 | 36.4 | 88.1 | 52.8 | 69.2 | 41.6 | — | 官方榜单 | 2026-10-16 | |
| 47 | GTA1-Qwen2.5VL-72B | — | — | — | 58.4 | 74.5 | 32.5 | 79.9 | 33.1 | 73.2 | 20.3 | 56.9 | 28.1 | 81.9 | 38.2 | 85.3 | 49.1 | 73.8 | 37.1 | — | 官方榜单 | 2026-10-16 | |
| 48 | Holo1.5-7B | — | — | — | 57.9 | 68.7 | 40.6 | 74.7 | 46.2 | 64.7 | 30.8 | 53.8 | 31.3 | 81.3 | 44.5 | 81.9 | 56.6 | 56.1 | 39.3 | — | 官方榜单 | 2026-10-16 | |
| 49 | BAMI-7B | — | — | — | 57.8 | 72.9 | 33.3 | 81.8 | 26.9 | 68.2 | 23.8 | 58.4 | 29.7 | 77.8 | 36.4 | 83.6 | 60.4 | 71.0 | 41.6 | — | 官方榜单 | 2026-10-16 | A test-time inference method for GUI grounding that uses coarse-to-fine focus and candidate selection. Base model: TianXi-Action-7B. |
| 49 | UI-Venus-1-5-2B | — | — | — | 57.8 | 69.1 | 39.4 | 70.1 | 43.5 | 63.6 | 28.7 | 54.3 | 32.8 | 76.4 | 38.2 | 81.9 | 47.2 | 73.8 | 51.7 | — | 官方榜单 | 2026-10-16 | |
| 51 | MAI-UI-2B | — | — | — | 57.4 | 73.5 | 31.3 | 76.6 | 32.4 | 69.2 | 21.7 | 61.4 | 23.4 | 81.3 | 34.5 | 85.9 | 39.6 | 68.2 | 41.6 | — | 官方榜单 | 2026-10-16 | |
| 52 | Holo2-4B | — | — | — | 57.2 | 70.0 | 36.6 | 79.2 | 39.3 | 63.6 | 25.9 | 44.2 | 31.3 | 80.6 | 36.4 | 85.3 | 54.7 | 76.6 | 42.7 | — | 官方榜单 | 2026-10-16 | |
| 53 | HuzzleWorld-1B | — | — | — | 55.6 | 68.2 | 35.3 | 76.0 | 40.0 | 68.7 | 25.2 | 52.8 | 26.6 | 77.1 | 36.4 | 75.1 | 43.4 | 60.8 | 43.8 | — | 官方榜单 | 2026-10-16 | Proprietary |
| 54 | GTA1-7B | — | — | — | 55.5 | 68.6 | 34.3 | 73.4 | 30.3 | 67.2 | 26.6 | 57.9 | 23.4 | 70.8 | 35.5 | 83.6 | 62.3 | 56.1 | 42.7 | — | 官方榜单 | 2026-10-16 | |
| 55 | GTA1-Qwen2.5VL-32B | — | — | — | 53.6 | 69.9 | 27.1 | 82.5 | 28.3 | 69.2 | 14.7 | 43.6 | 23.4 | 79.9 | 31.8 | 80.8 | 43.4 | 70.1 | 32.6 | — | 官方榜单 | 2026-10-16 | |
| 56 | Qwen2.5-VL-72B-Instruct | — | — | — | 53.3 | 69.4 | 27.3 | 78.6 | 26.9 | 62.6 | 20.3 | 54.3 | 14.1 | 77.8 | 34.5 | 80.2 | 47.2 | 67.3 | 28.1 | — | 官方榜单 | 2026-10-16 | |
| 57 | GUI-Spotlight | — | — | — | 52.8 | 71.4 | 22.7 | 74.1 | 21.9 | 63.9 | 21.2 | 67.1 | 17.7 | 77.8 | 21.8 | 80.8 | 35.9 | 63.5 | 24.7 | — | 官方榜单 | 2026-10-16 | A think-with-image GUI visual grounding model. |
| 58 | V2P-7B | — | — | — | 52.5 | 69.4 | 25.2 | 72.7 | 20.7 | 61.1 | 16.8 | 63.5 | 12.5 | 77.1 | 31.8 | 79.7 | 50.9 | 63.5 | 31.5 | — | 官方榜单 | 2026-10-16 | |
| 59 | TianXi-Action-7B | — | — | — | 51.9 | 67.0 | 27.3 | 76.0 | 21.4 | 61.6 | 19.6 | 45.2 | 18.8 | 80.6 | 31.8 | 84.2 | 54.7 | 57.9 | 33.7 | — | 官方榜单 | 2026-10-16 | |
| 60 | Holo1.5-3B | — | — | — | 51.5 | 64.9 | 29.8 | 73.4 | 28.3 | 63.1 | 22.4 | 50.3 | 34.4 | 72.9 | 40.0 | 78.0 | 34.0 | 50.5 | 25.8 | — | 官方榜单 | 2026-10-16 | |
| 61 | Kimi-VL-A3B-Thinking-2506 | — | — | — | 51.0 | 66.8 | 25.3 | 67.5 | 18.6 | 62.1 | 20.3 | 66.0 | 17.2 | 74.3 | 30.0 | 75.1 | 41.5 | 52.3 | 34.8 | — | 官方榜单 | 2026-10-16 | |
| 62 | UI-Venus-7B | — | — | — | 50.8 | 67.1 | 24.3 | 74.7 | 24.1 | 63.1 | 14.7 | 60.4 | 21.9 | 76.4 | 31.8 | 75.7 | 41.5 | 49.5 | 22.5 | — | 官方榜单 | 2026-10-16 | |
| 63 | GTA1-Qwen2.5VL-7B | — | — | — | 50.1 | 65.5 | 25.2 | 66.9 | 20.7 | 62.6 | 18.2 | 53.3 | 17.2 | 76.4 | 31.8 | 82.5 | 50.9 | 48.6 | 25.8 | — | 官方榜单 | 2026-10-16 | |
| 64 | Qwen2.5-VL-32B-Instruct | — | — | — | 48.0 | 62.6 | 24.3 | 74.0 | 22.1 | 61.1 | 16.1 | 34.5 | 20.3 | 75.0 | 30.0 | 74.6 | 30.2 | 64.5 | 33.7 | — | 官方榜单 | 2026-10-16 | |
| 65 | UI-AGILE-7B | — | — | — | 47.9 | 63.8 | 22.2 | 66.2 | 24.1 | 59.6 | 10.5 | 52.8 | 12.5 | 66.7 | 26.4 | 78.5 | 37.7 | 59.8 | 30.3 | — | 官方榜单 | 2026-10-16 | UI-AGILE enhances GUI agents through improved training with a Continuous Reward function, and Cropping-based Resampling, and inference with Decomposed Grounding with Selection. |
| 66 | SE-GUI-7B | — | — | — | 47.3 | 63.5 | 21.0 | 68.2 | 19.3 | 57.6 | 9.1 | 51.3 | 14.1 | 75.0 | 28.2 | 78.5 | 43.4 | 49.5 | 25.8 | — | 官方榜单 | 2026-10-16 | |
| 67 | UI-AGILE-3B | — | — | — | 45.0 | 61.2 | 18.7 | 66.9 | 16.6 | 58.1 | 11.9 | 47.2 | 10.9 | 66.7 | 24.6 | 72.3 | 34.0 | 58.9 | 22.5 | — | 官方榜单 | 2026-10-16 | UI-AGILE enhances GUI agents through improved training with a Continuous Reward function, and Cropping-based Resampling, and inference with Decomposed Grounding with Selection. |
| 68 | GUI-Actor-2.5VL-7B | — | — | — | 44.6 | 60.0 | 19.7 | 59.1 | 15.9 | 59.6 | 16.1 | 47.7 | 9.4 | 70.1 | 25.4 | 69.5 | 41.5 | 55.1 | 19.1 | — | 官方榜单 | 2026-10-16 | |
| 69 | GUI-Actor-2.5VL-3B | — | — | — | 42.2 | 56.9 | 18.4 | 65.6 | 12.4 | 52.0 | 14.0 | 40.6 | 14.1 | 68.1 | 25.4 | 70.1 | 32.1 | 46.7 | 21.4 | — | 官方榜单 | 2026-10-16 | |
| 70 | GUI-Actor-2VL-7B | — | — | — | 40.7 | 55.8 | 16.4 | 60.4 | 15.9 | 60.1 | 12.6 | 36.0 | 9.4 | 63.2 | 20.0 | 64.4 | 30.2 | 53.3 | 15.7 | — | 官方榜单 | 2026-10-16 | |
| 71 | UI-TARS-72B | — | — | — | 38.1 | 50.9 | 17.6 | 63.0 | 17.2 | 57.1 | 15.4 | 18.8 | 12.5 | 64.6 | 20.9 | 63.3 | 26.4 | 42.1 | 15.7 | — | 官方榜单 | 2026-10-16 | |
| 72 | GUI-Actor-2VL-2B | — | — | — | 36.7 | 50.7 | 14.1 | 55.8 | 12.4 | 54.5 | 9.1 | 34.5 | 9.4 | 54.2 | 17.3 | 62.7 | 24.5 | 41.1 | 18.0 | — | 官方榜单 | 2026-10-16 | |
| 73 | GUI-G1 | — | — | — | 35.9 | 48.3 | 15.9 | 51.3 | 12.4 | 38.4 | 11.9 | 39.1 | 7.8 | 58.3 | 27.3 | 66.7 | 32.1 | 35.5 | 10.1 | — | 官方榜单 | 2026-10-16 | |
| 74 | SE-GUI-3B | — | — | — | 35.9 | 50.4 | 11.9 | 58.4 | 10.3 | 46.0 | 5.6 | 38.1 | 12.5 | 61.8 | 19.1 | 59.3 | 22.6 | 39.3 | 9.0 | — | 官方榜单 | 2026-10-16 | |
| 75 | UI-TARS-7B | — | — | — | 35.7 | 47.8 | 16.2 | 58.4 | 12.4 | 50.0 | 9.1 | 20.8 | 9.4 | 63.9 | 31.8 | 63.3 | 20.8 | 30.8 | 16.9 | — | 官方榜单 | 2026-10-16 | |
| 76 | UGround-v1-72B | — | — | — | 34.5 | 49.0 | 11.1 | 55.8 | 4.8 | 54.0 | 10.5 | 16.8 | 4.7 | 70.8 | 22.7 | 61.0 | 18.9 | 40.2 | 7.9 | — | 官方榜单 | 2026-10-16 | |
| 77 | Di-GUI-3B | — | — | — | 31.6 | 44.2 | 11.1 | 34.4 | 7.6 | 49.0 | 5.6 | 33.0 | 12.5 | 51.4 | 20.0 | 63.3 | 18.9 | 29.0 | 9.0 | — | 官方榜单 | 2026-10-16 | |
| 78 | UGround-V1-7B | — | — | — | 31.1 | 45.2 | 8.1 | 52.0 | 2.8 | 47.5 | 9.8 | 18.8 | 1.6 | 57.6 | 14.6 | 60.5 | 13.2 | 38.3 | 7.9 | — | 官方榜单 | 2026-10-16 | |
| 79 | ZonUI-3B | — | — | — | 28.7 | 39.2 | 11.8 | 24.7 | 6.2 | 40.9 | 7.7 | 32.0 | 15.6 | 54.9 | 18.2 | 57.1 | 26.4 | 19.6 | 7.9 | — | 官方榜单 | 2026-10-16 | |
| 80 | UI-TARS-2B | — | — | — | 27.7 | 39.6 | 8.4 | 47.4 | 4.1 | 42.9 | 6.3 | 17.8 | 4.7 | 56.9 | 17.3 | 50.3 | 17.0 | 21.5 | 5.6 | — | 官方榜单 | 2026-10-16 | |
| 81 | Qwen2.5-VL-7B-Instruct | — | — | — | 26.8 | 38.9 | 7.1 | 46.8 | 4.1 | 35.9 | 7.7 | 16.8 | 1.6 | 49.3 | 7.3 | 52.5 | 20.8 | 37.4 | 6.7 | — | 官方榜单 | 2026-10-16 | |
| 82 | UGround-V1-2B | — | — | — | 26.6 | 38.8 | 7.0 | 50.0 | 3.5 | 42.9 | 4.2 | 17.3 | 6.3 | 50.0 | 13.6 | 45.8 | 13.2 | 28.0 | 5.6 | — | 官方榜单 | 2026-10-16 | |
| 83 | Aguvis-7B | — | — | — | 22.9 | 33.9 | 5.1 | 30.5 | 0.7 | 34.9 | 2.8 | 17.3 | 3.1 | 52.8 | 10.9 | 40.1 | 15.1 | 31.8 | 4.5 | — | 官方榜单 | 2026-10-16 | |
| 84 | OS-Atlas-7B | — | — | — | 18.9 | 28.1 | 4.0 | 33.1 | 1.4 | 28.8 | 2.8 | 12.2 | 4.7 | 37.5 | 7.3 | 33.9 | 5.7 | 27.1 | 4.5 | — | 官方榜单 | 2026-10-16 | |
| 85 | GPT5-minimal (resized) | — | — | — | 18.5 | 26.4 | 5.8 | 22.1 | 2.1 | 28.8 | 4.2 | 17.3 | 4.7 | 41.7 | 10.0 | 27.7 | 7.5 | 22.4 | 9.0 | — | 官方榜单 | 2026-10-16 | |
| 86 | Claude (Computer Use) | — | — | — | 17.1 | 23.4 | 7.1 | 22.0 | 3.9 | 25.9 | 3.4 | 14.6 | 3.7 | 33.9 | 15.8 | 30.1 | 16.3 | 11.0 | 4.5 | — | 官方榜单 | 2026-10-16 | |
| 87 | UGround-7B | — | — | — | 16.5 | 25.0 | 2.8 | 26.6 | 2.1 | 27.3 | 2.8 | 14.2 | 1.6 | 31.9 | 2.7 | 31.6 | 11.3 | 17.8 | 0.0 | — | 官方榜单 | 2026-10-16 | |
| 88 | Qwen2.5-VL-3B-Instruct | — | — | — | 16.1 | 23.6 | 3.8 | 22.1 | 1.4 | 26.8 | 2.1 | 9.1 | 1.6 | 38.2 | 7.3 | 33.9 | 15.1 | 10.3 | 1.1 | — | 官方榜单 | 2026-10-16 | |
| 89 | AriaUI (MOE, 3.9B active) | — | — | — | 11.3 | 17.1 | 2.0 | 16.2 | 0.0 | 23.7 | 2.1 | 7.6 | 1.6 | 27.1 | 6.4 | 20.3 | 1.9 | 4.7 | 0.0 | — | 官方榜单 | 2026-10-16 | |
| 90 | CogAgent (18B) | — | — | — | 7.7 | 12.0 | 0.8 | 14.9 | 0.7 | 9.6 | 0.0 | 7.1 | 3.1 | 22.2 | 1.8 | 13.0 | 0.0 | 5.6 | 0.0 | — | 官方榜单 | 2026-10-16 | |
| 90 | ShowUI (2B) | — | — | — | 7.7 | 10.8 | 2.6 | 16.9 | 1.4 | 9.1 | 0.0 | 2.5 | 0.0 | 13.2 | 7.3 | 15.3 | 7.5 | 10.3 | 2.3 | — | 官方榜单 | 2026-10-16 | |
| 92 | GPT5-high (resized) | — | — | — | 6.0 | 7.8 | 3.1 | 6.5 | 1.4 | 10.6 | 2.8 | 6.1 | 4.7 | 11.8 | 5.5 | 5.7 | 7.5 | 5.6 | 0.0 | — | 官方榜单 | 2026-10-16 | |
| 93 | OS-Atlas-4B | — | — | — | 3.7 | 5.0 | 1.7 | 7.1 | 0.0 | 3.0 | 1.4 | 2.0 | 0.0 | 9.0 | 5.5 | 5.1 | 3.8 | 5.6 | 0.0 | — | 官方榜单 | 2026-10-16 | |
| 94 | MiniCPM-V (7B) | — | — | — | 3.0 | 4.5 | 0.7 | 7.1 | 0.0 | 2.0 | 0.0 | 4.1 | 1.6 | 8.3 | 0.0 | 2.8 | 3.8 | 3.7 | 1.1 | — | 官方榜单 | 2026-10-16 | |
| 95 | Qwen2-VL-7B | — | — | — | 1.6 | 2.5 | 0.2 | 2.6 | 0.0 | 1.5 | 0.0 | 0.5 | 0.0 | 6.3 | 0.0 | 3.4 | 1.9 | 0.9 | 0.0 | — | 官方榜单 | 2026-10-16 | |
| 96 | SeeClick (7B) | — | — | — | 1.1 | 1.8 | 0.0 | 0.7 | 0.0 | 1.0 | 0.0 | 2.5 | 0.0 | 3.5 | 0.0 | 1.1 | 0.0 | 2.8 | 0.0 | — | 官方榜单 | 2026-10-16 | |
| 97 | Qwen2-VL-72B-Instruct | — | — | — | 1.0 | 1.4 | 0.3 | 1.9 | 0.0 | 1.0 | 0.0 | 1.0 | 0.0 | 4.2 | 0.0 | 0.6 | 1.9 | 0.0 | 1.1 | — | 官方榜单 | 2026-10-16 | |
| 98 | GPT-4o | — | — | — | 0.8 | 1.3 | 0.0 | 1.3 | 0.0 | 1.0 | 0.0 | 2.0 | 0.0 | 2.1 | 0.0 | 1.1 | 0.0 | 0.0 | 0.0 | — | 官方榜单 | 2026-10-16 | |
| 99 | Qwen-VL-7B | — | — | — | 0.1 | 0.1 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.7 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | — | 官方榜单 | 2026-10-16 | |
榜单说明
第 1 批接入- 指标
- 定位准确率：定位正确的样本占比（%），榜单注明按样本微平均（越高越好）
- 子项
- 总体TextIcon开发TextIcon创意TextIconCADTextIcon科学TextIcon办公TextIcon操作系统TextIcon
- 数据源
- 来源类型
- 官方榜单 98模型方自报 1
分数取自数据源原文；同一模型在不同评测设置或不同来源下分别列出，来源冲突时全部保留并标注来源类型。
更新记录
0 条首次同步作为基线，之后新上榜、分数变化和移除的条目会记录在这里。