multi-step operation sequence prediction on Android-In-The-Zoo
74.43PrecisionTeach VLM
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Teach VLM2026.06 | 74.43 | 73.85 | 74.14 | 25.56 | |
| Doubao-Seed-1.82026.06 | 44.13 | 35.56 | 39.39 | 55.87 | |
| Doubao-Seed-2.02026.06 | 41.14 | 36.03 | 38.42 | 58.86 | |
| Qwen3-VL-8B2026.06 | 34.62 | 25.7 | 29.5 | 65.38 | |
| GPT-4o2026.06 | 19.15 | 16.7 | 17.84 | 80.85 | |
| Qwen3.5-9B2026.06 | 5.79 | 13.27 | 8.06 | 94.21 |