Clarification-aware multi-agent orchestration is challenging because systems must elicit missing information, maintain task state, and coordinate dependent subtasks. Existing benchmarks do not comprehensively evaluate such orchestration under underspecified user requests, focusing instead on end-to-end success or isolated abilities like tool use and dialogue. We introduce MACRO, a travel-domain benchmark designed to address these gaps by providing structured task tiers, process-level user-side and system-side milestone annotations that include expected user clarifications, and standardized sub-agent interfaces. MACRO-Full contains 107 fully annotated tasks with 703 milestones, and MACRO-Core provides a tractable 30-task subset with 164 milestones. We evaluate three representative orchestration frameworks and a reference decoupled planning–execution orchestrator (MACRO-DPE) using trace-level metrics and failure analysis. Results show that MACRO-DPE improves routing fidelity and execution efficiency while maintaining task completion, whereas staged planning ensures milestone coverage at higher cost. The dataset[<https://huggingface.co/datasets/weier248/MACRO-bench>] and code are publicly available at https://github.com/liuweie/MACRO-bench/.

Generating well-structured long music compositions, spanning several minutes, remains a challenge due to inefficient representation and the lack of structured representation. In this paper, we propose PyramidCodec, a hierarchical discrete representation of audio, for long audio-domain music generation. Specifically, we employ residual vector quantization on different levels of features to obtain the hierarchical discrete representation. The highest level of features has the largest hop size, resulting in the most compact token sequence. The quantized higher-level representation is up-sampled and combined with lower-level features to apply residual vector quantization and obtain lower-level discrete representations. Furthermore, we design a hierarchical training strategy to ensure that the details are gradually added with more levels of tokens. By performing hierarchical tokenization, the overall token sequence represents information at various scales, facilitating long-context modeling in music and enabling the generation of well-structured compositions. The experimental results demonstrate that our proposed PyramidCodec achieves competitive performance in terms of reconstruction quality and token per second (TPS). By enabling ultra-long music modeling at the lowest level, the proposed approach facilitates training a language model that can generate well-structured long-form music for up to 3 minutes, whose quality is further demonstrated by subjective and objective evaluations. The samples can be found at https://pyramidcodec.github.io/.