Operational Inventory Management on RetailBench Middle 20 Categories
29.33Avg DaysQwen-235B
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Qwen-235B2026.03 | 29.33 | 216.06 | 179.42 | 6.39 | 13.33 | 45 | |
| Kimi-K2variant=Thinking2026.03 | 37 | 347.78 | 356.1 | 0.37 | 16.77 | 50 | |
| Gemini-3variant=Fast2026.03 | 42.67 | 439.05 | 449.04 | 1.88 | 16.3 | 48 | |
| Average (7 models)2026.03 | 46.48 | 364.24 | 282.48 | 4.91 | 13.99 | 54 | |
| Grok-4.1 Fast2026.03 | 51 | 720.69 | 398.23 | 4.07 | 11.6 | 59 | |
| GLM-4.62026.03 | 54.33 | 182.55 | 131.9 | 2.37 | 12.74 | 56 | |
| DeepSeek-V3.2variant=Exp.2026.03 | 54.67 | 263.68 | 255.3 | 10.64 | 18.18 | 63 | |
| OpenAI-5.1 Mini2026.03 | 56.33 | 336.24 | 223.6 | 13.07 | 12.35 | 57 | |
| Hand-crafted Policy2026.03 | 180 | 1,870.21 | 2,809.39 | 2.72 | 0.74 | 180 |