Visual Tool-use on GTA 121-case (val)
86.4Tool AccuracyDynamo
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| DynamoModel=GPT-5.42026.06 | 86.4 | 63.4 | 63.1 | 74.4 | |
| DynamoModel=Doubao-Seed-2.02026.06 | 86.4 | 61.3 | 57.7 | 81.8 | |
| DynamoModel=GPT-4o2026.06 | 84.6 | 64.2 | 63.4 | 71.1 | |
| Base AgentModel=GPT-4o2026.06 | 80.3 | 62 | 52 | 66.1 | |
| Base AgentModel=GPT-5.42026.06 | 79.6 | 60.2 | 57 | 72.7 | |
| XSkill-styleModel=Doubao-Seed-2.02026.06 | 72.8 | 48.4 | 44.4 | 76.9 | |
| XSkill-styleModel=GPT-4o2026.06 | 71.3 | 53 | 49.5 | 68.6 | |
| Base AgentModel=Doubao-Seed-2.02026.06 | 67.7 | 45.5 | 41.6 | 76.9 | |
| XSkill-styleModel=GPT-5.42026.06 | 66.3 | 48 | 46.2 | 71.1 |