Operational Inventory Management on RetailBench Easy 5 Categories
37.5Avg. DaysQwen-235B
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Qwen-235B2026.03 | 37.5 | 420.31 | 236.5 | 7.45 | 8.41 | 48 | |
| Gemini-3variant=Fast2026.03 | 50.67 | 399.39 | 294.71 | 7.99 | 13.11 | 59 | |
| OpenAI-5.1 Mini2026.03 | 51.75 | 192.9 | 122.46 | 3.6 | 12.37 | 55 | |
| Average (7 models)2026.03 | 52.38 | 301.98 | 203.3 | 5.9 | 10.68 | 61.71 | |
| GLM-4.62026.03 | 52.4 | 174.34 | 124.67 | 7.73 | 12.93 | 58 | |
| Kimi-K2variant=Thinking2026.03 | 54.25 | 260.68 | 168.72 | 2.39 | 11.79 | 58 | |
| DeepSeek-V3.2variant=Exp.2026.03 | 58.33 | 229.19 | 183.26 | 8.89 | 11.22 | 66 | |
| Grok-4.1 Fast2026.03 | 61.75 | 508.08 | 336.94 | 4.17 | 8.47 | 88 | |
| Hand-crafted Policy2026.03 | 180 | 674.18 | 729.46 | 2.66 | 0.7 | 180 |