ResearchBenchmarksMulti-modal Agent Reasoning on GAIA (full)Follow38.8Pass@1Reagent-U19.24824.32429.434.476Jan 29, 2026Evaluation ResultsMethodMethodLinksPass@1Pass@3Reagent-U2026.0138.853.9MCP-R12026.0137.651.5Qwen3-8BModel size=8BModel size=8B2026.012026.7