E-commerce shopping agent on ShoppingBench
60.8Products Finder ASRSFT+RL-Qwen3-4B
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| SFT+RL-Qwen3-4BModel Parameters=4B, Training Strategy=SFT+RL2025.08 | 60.8 | 86.1 | 46.7 | 51.3 | 53.2 | 85.5 | 33.2 | 79 | 48.7 | |
| GPT-4.1Model Accessibility=Closed-Source2025.08 | 59.6 | 83.6 | 62 | 67.3 | 46.4 | 79.2 | 30.4 | 72.8 | 48.2 | |
| Qwen2.5-maxModel Accessibility=Closed-Source2025.08 | 58.4 | 81 | 42.7 | 50.7 | 22.8 | 67.3 | 22.4 | 65.8 | 35.9 | |
| SFT-Qwen3-4BModel Parameters=4B, Training Strategy=SFT2025.08 | 55.6 | 81.1 | 52.7 | 59.3 | 39.2 | 77.9 | 30.4 | 76 | 43.6 | |
| DeepSeek-V3Model Accessibility=Open-Source2025.08 | 54.8 | 75.8 | 48 | 54.7 | 22.8 | 46.9 | 21.2 | 55.3 | 35.4 | |
| DeepSeek-R1Model Accessibility=Open-Source2025.08 | 53.2 | 75.8 | 44 | 53.3 | 37.2 | 51.5 | 24.4 | 43.9 | 39.2 | |
| GPT-4oModel Accessibility=Closed-Source2025.08 | 52.4 | 71.5 | 50 | 58.7 | 24 | 52.4 | 25.2 | 65.6 | 36.6 | |
| Qwen3-32BModel Accessibility=Open-Source, Model Parameters=32B2025.08 | 51.6 | 77.6 | 45.3 | 54 | 25.6 | 54.2 | 18 | 63.8 | 34 | |
| Gemini-2.5-FlashModel Accessibility=Closed-Source2025.08 | 49.2 | 71.3 | 39.3 | 46.7 | 32 | 40.9 | 22.8 | 55 | 35.4 | |
| Qwen3-235B-A22BModel Accessibility=Open-Source, Model Parameters=235B2025.08 | 49.2 | 77.2 | 40 | 46.7 | 28.8 | 56.3 | 14.4 | 55.3 | 32.3 | |
| Claude-4-SonnetModel Accessibility=Closed-Source2025.08 | 48 | 73.1 | 51.3 | 62.7 | 37.6 | 59.5 | 24 | 72.2 | 39 | |
| Qwen3-14BModel Accessibility=Open-Source, Model Parameters=14B2025.08 | 46 | 70.4 | 30 | 37.3 | 19.2 | 53.1 | 12.4 | 58.1 | 26.6 | |
| o3-miniModel Accessibility=Closed-Source2025.08 | 42 | 62.6 | 51.3 | 57.3 | 36.8 | 50.3 | 31.6 | 61.4 | 39.2 | |
| Qwen3-8BModel Accessibility=Open-Source, Model Parameters=8B2025.08 | 40 | 65.8 | 22.7 | 27.3 | 13.6 | 31.7 | 11.2 | 53.2 | 21.8 | |
| Qwen3-4BModel Accessibility=Open-Source, Model Parameters=4B2025.08 | 36.4 | 66.4 | 18.7 | 26 | 8.8 | 29.8 | 8.4 | 45.5 | 18 | |
| GPT-4o-miniModel Accessibility=Closed-Source2025.08 | 33.2 | 46.9 | 28 | 31.3 | 10.4 | 52.7 | 11.6 | 54.6 | 20 | |
| Gemma-3-27BModel Accessibility=Open-Source, Model Parameters=27B2025.08 | 32 | 48.5 | 46.7 | 57.3 | 18 | 65.4 | 17.2 | 62.5 | 26.5 | |
| Gemma-3-12BModel Accessibility=Open-Source, Model Parameters=12B2025.08 | 27.2 | 42.5 | 32 | 36.7 | 9.6 | 51.7 | 13.6 | 55.7 | 19.3 | |
| Gemma-3-4BModel Accessibility=Open-Source, Model Parameters=4B2025.08 | 24.4 | 40.1 | 16.7 | 20.7 | 0 | 31.1 | 4.8 | 30.7 | 10.9 |