Multi-Hop Question Answering on HotpotQA (Pass@1)
60.8Pass@1AFLOW
Evaluation Results
| Method | Links | |
|---|---|---|
| AFLOWBackbone=GPT-5-Chat2026.04 | 60.8 | |
| Mem2EvolveBackbone=GPT-5-Chat2026.04 | 60.8 | |
| AlitaBackbone=GPT-5-Chat2026.04 | 58.8 | |
| SwarmAgenticBackbone=GPT-5-Chat2026.04 | 56 | |
| DSPyBackbone=GPT-5-Chat2026.04 | 55.6 | |
| EvoAgentBackbone=GPT-5-Chat2026.04 | 54.4 | |
| AutoAgentsBackbone=GPT-5-Chat2026.04 | 54.2 | |
| DyLANBackbone=GPT-5-Chat2026.04 | 52 | |
| GPT-5-Chat (Direct)Backbone=GPT-5-Chat2026.04 | 50.4 | |
| GPT-5-Chat (CoT)Backbone=GPT-5-Chat2026.04 | 47.4 | |
| GPT-5-Chat (ReAct)Backbone=GPT-5-Chat2026.04 | 41.4 | |
| AgentVerseBackbone=GPT-5-Chat2026.04 | 38.6 |