Target Product Selection on Target Product Selection transfer to fine-tunes
95.5Selection Success RatePRAC
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PRACOptim Model=Qwen3-VL, Transfer Model=SEA-LION-VL, Agent Setting=ReAct-S2026.04 | 95.5 | 0.2 | |
| PRACOptim Model=Qwen3-VL, Transfer Model=PEARL, Agent Setting=ReAct-S2026.04 | 93.3 | 2 | |
| PRACOptim Model=Qwen3-VL, Transfer Model=SEA-LION-VL, Agent Setting=Action2026.04 | 85.8 | 3.3 | |
| PRACOptim Model=Qwen3-VL, Transfer Model=PEARL, Agent Setting=Action2026.04 | 81.5 | 1 | |
| PRACOptim Model=GLM4.6V‡, Transfer Model=GLM-4.6V-FDM, Agent Setting=ReAct-S2026.04 | 80.3 | 0.2 | |
| PRACOptim Model=GLM4.6V‡, Transfer Model=GLM-4.6V-FDM, Agent Setting=Action2026.04 | 77.9 | 0.2 | |
| CE-tOptim Model=Qwen3-VL, Transfer Model=SEA-LION-VL, Agent Setting=ReAct-S2026.04 | 67 | 3 | |
| CE-tOptim Model=Qwen3-VL, Transfer Model=PEARL, Agent Setting=ReAct-S2026.04 | 66.7 | 3.3 | |
| CE-tOptim Model=GLM4.6V‡, Transfer Model=GLM-4.6V-FDM, Agent Setting=ReAct-S2026.04 | 65.5 | 0.6 | |
| CE-tOptim Model=GLM4.6V‡, Transfer Model=GLM-4.6V-FDM, Agent Setting=Action2026.04 | 62.1 | 1 | |
| PRACOptim Model=Qwen3-VL, Transfer Model=Cosmos-Reason2, Agent Setting=ReAct-S2026.04 | 62 | 33.3 | |
| TOOptim Model=Qwen3-VL, Transfer Model=PEARL, Agent Setting=ReAct-S2026.04 | 57 | 2 | |
| TOOptim Model=Qwen3-VL, Transfer Model=SEA-LION-VL, Agent Setting=ReAct-S2026.04 | 49.9 | 5.1 | |
| TOOptim Model=GLM4.6V‡, Transfer Model=GLM-4.6V-FDM, Agent Setting=ReAct-S2026.04 | 41.2 | 0.6 | |
| CE-tOptim Model=Qwen3-VL, Transfer Model=SEA-LION-VL, Agent Setting=Action2026.04 | 40.2 | 5.9 | |
| TOOptim Model=GLM4.6V‡, Transfer Model=GLM-4.6V-FDM, Agent Setting=Action2026.04 | 39.4 | 0.4 | |
| PRACOptim Model=Qwen3-VL, Transfer Model=Cosmos-Reason2, Agent Setting=Action2026.04 | 39.3 | 43.2 | |
| CE-tOptim Model=Qwen3-VL, Transfer Model=Cosmos-Reason2, Agent Setting=ReAct-S2026.04 | 37.2 | 32.8 | |
| CE-tOptim Model=Qwen3-VL, Transfer Model=PEARL, Agent Setting=Action2026.04 | 36.5 | 2.2 | |
| TOOptim Model=Qwen3-VL, Transfer Model=Cosmos-Reason2, Agent Setting=ReAct-S2026.04 | 34 | 21 | |
| TOOptim Model=Qwen3-VL, Transfer Model=PEARL, Agent Setting=Action2026.04 | 28.1 | 3.8 | |
| TOOptim Model=Qwen3-VL, Transfer Model=SEA-LION-VL, Agent Setting=Action2026.04 | 27.9 | 3.6 | |
| TOOptim Model=Qwen3-VL, Transfer Model=Cosmos-Reason2, Agent Setting=Action2026.04 | 26.5 | 2.2 | |
| CE-tOptim Model=Qwen3-VL, Transfer Model=Cosmos-Reason2, Agent Setting=Action2026.04 | 26.2 | 8.1 | |
| CleanOptim Model=GLM4.6V‡, Transfer Model=GLM-4.6V-FDM, Agent Setting=Action2026.04 | 22.2 | 0.8 | |
| CleanOptim Model=GLM4.6V‡, Transfer Model=GLM-4.6V-FDM, Agent Setting=ReAct-S2026.04 | 22 | 1.3 | |
| CleanOptim Model=Qwen3-VL, Transfer Model=PEARL, Agent Setting=ReAct-S2026.04 | 20.9 | 0.1 | |
| CleanOptim Model=Qwen3-VL, Transfer Model=Cosmos-Reason2, Agent Setting=ReAct-S2026.04 | 20.4 | 0.4 | |
| CleanOptim Model=Qwen3-VL, Transfer Model=SEA-LION-VL, Agent Setting=Action2026.04 | 20.2 | 0 | |
| CleanOptim Model=Qwen3-VL, Transfer Model=Cosmos-Reason2, Agent Setting=Action2026.04 | 20 | 0.2 | |
| CleanOptim Model=Qwen3-VL, Transfer Model=PEARL, Agent Setting=Action2026.04 | 19.8 | 0.4 | |
| CleanOptim Model=Qwen3-VL, Transfer Model=SEA-LION-VL, Agent Setting=ReAct-S2026.04 | 19.7 | 1.1 |