Multi-hop verification on HoVer
1,255LLM Throughput (token/s)FlashEvolve
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FlashEvolveDeployment=vLLM, Model=Qwen3-8B2026.05 | 1,255 | 5.9 | |
| CombeeDeployment=vLLM, Model=Qwen3-8B, K (parallelization size)=402026.05 | 891 | 2 | |
| CombeeDeployment=vLLM, Model=Qwen3-8B, K (parallelization size)=102026.05 | 810 | 2 | |
| GEPADeployment=vLLM, Model=Qwen3-8B2026.05 | 461 | 2.5 | |
| FlashEvolveDeployment=API, Model=GPT-4o-mini2026.05 | 352 | 9.1 | |
| CombeeDeployment=API, Model=GPT-4o-mini, K (parallelization size)=102026.05 | 348 | 0.8 | |
| CombeeDeployment=API, Model=GPT-4o-mini, K (parallelization size)=402026.05 | 214 | 0.7 | |
| GEPADeployment=API, Model=GPT-4o-mini2026.05 | 142 | 1.8 |