Multi-hop reasoning and question answering on GAIA
72GAIA L1 AccAgentConductor
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| AgentConductorBackbone=3B2026.02 | 72 | 53.4 | 36.1 | 53.8 | |
| Chain-of-AgentsBackbone=7B2026.02 | 69.2 | 50.9 | 33.3 | 50.8 |
Loading the SOTA2 catalog…
| Method | Links | ||||
|---|---|---|---|---|---|
| AgentConductorBackbone=3B2026.02 | 72 | 53.4 | 36.1 | 53.8 | |
| Chain-of-AgentsBackbone=7B2026.02 | 69.2 | 50.9 | 33.3 | 50.8 |