Subway Scheduling on NYC Manhattan (out-of-domain)
250.58Electricity ConsumptionKimi-K2.5
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Kimi-K2.5Framework=TrafficClaw2026.04 | 250.58 | 357.27 | |
| Qwen3-32BFramework=TrafficClaw2026.04 | 258.85 | 412.06 | |
| DeepSeek-V3.2Framework=TrafficClaw2026.04 | 260.82 | 358.27 | |
| Qwen3-235BFramework=TrafficClaw2026.04 | 270.22 | 427.45 | |
| AIDE2026.04 | 276.31 | 425.14 | |
| Gemini-3.1-ProFramework=TrafficClaw2026.04 | 287.9 | 387.61 | |
| MiniMax-M2.5Framework=TrafficClaw2026.04 | 291 | 324.92 | |
| TrafficClaw (8B)Model size=8B2026.04 | 294.53 | 220.99 | |
| TrafficGPT2026.04 | 314.76 | 678.61 | |
| SUMO-MCP2026.04 | 315.41 | 294.61 | |
| Qwen3-8BFramework=TrafficClaw2026.04 | 317.08 | 350.92 | |
| Classic Method2026.04 | 330.04 | 603.01 | |
| o4-miniFramework=TrafficClaw2026.04 | 352.94 | 259.82 |