List your company

Multi-step reasoning and tool use on GAIA (val)

80Pass@1AORCHESTRA
39.034449.669760.30570.9403Feb 3, 2026

Evaluation results

MethodLinks
AORCHESTRAModel Setup=Gemini-3-Flash2026.028086.0671.62
AORCHESTRAModel Setup=DeepSeek-V3.22026.0267.878058.43
OpenHandsModel Setup=Gemini-3-Flash2026.0266.0672.7348.49
OpenHandsModel Setup=DeepSeek-V3.22026.0263.6472.1248.35
AORCHESTRAModel Setup=Claude-4.5-haiku2026.0260.6173.955.44
Mini-SWEModel Setup=Gemini-3-Flash2026.0258.1868.4849.49
OpenHandsModel Setup=Claude-4.5-haiku2026.0254.5561.2145.13
Mini-SWEModel Setup=DeepSeek-V3.22026.0250.363.6354.76
ReActModel Setup=Gemini-3-Flash2026.0249.0966.0647.22
ReActModel Setup=Claude-4.5-haiku2026.0247.8862.4243.62
ReActModel Setup=DeepSeek-V3.22026.0246.771.5138.23
Mini-SWEModel Setup=Claude-4.5-haiku2026.0240.616036.3