List your company

Multi-hop Question Answering on HotPotQA (Acc, TTFT, Delay (s))

60.3AccuracyStreamingThinker
53.43655.2185758.782Oct 20, 2025

Evaluation results

MethodLinks
StreamingThinkerParadigm=Streaming, Reasoning Depth=D3, Model=Qwen3-4B2025.1060.324.326.5
InterleavedParadigm=Interleaved, Reasoning Depth=D3, Model=Qwen3-4B2025.1059.824.3214.45
StreamingThinkerParadigm=Streaming, Reasoning Depth=D2, Model=Qwen3-4B2025.1058.124.323.92
InterleavedParadigm=Interleaved, Reasoning Depth=D2, Model=Qwen3-4B2025.1057.624.3212.02
BatchParadigm=Batch, Model=Qwen3-4B2025.1057.51,485.54720.37
StreamingThinkerParadigm=Streaming, Reasoning Depth=D1, Model=Qwen3-4B2025.1054.424.320.69
InterleavedParadigm=Interleaved, Reasoning Depth=D1, Model=Qwen3-4B2025.1053.724.328.33