- Rating
- 2
- Confidence
- 4
- Informativeness
- 1
- Presentation
- 3
- Soundness
- 3
Read review
Paper summary
Orchestra integrates instruction decomposition, grounding, visual history, and learned subtask transitions around an adapted pi_0.5 policy. Its contribution is primarily this particular integration, the policy adaptations, and RMBench-Compose for evaluating longer compositions after atomic-skill training.
Critical strengths
- Strong empirical results:
- Orchestra achieves 68.22% RMBench success (vs. Mem-0's 42%)
- Orchestra achieves 7.35 Average Success Length (ASL) on RMBench-Compose (vs. HarnessVLA's 2.986).
- Real-world mean ASL improves from 2.20 to 2.75 over Orchestra with the base policy.
- Implementation and evaluation contributions:
- The policy combines spatial prompts with temporal visual fusion without increasing the language model's visual-token count.
- Component ablations are useful and learned transitions match oracle transitions in aggregate ASL.
Critical weaknesses
- Distinct contribution over existing hierarchies
- Hierarchical instruction following and success-based subtask switching already appear in Hi Robot and Gemini Robotics 1.5.
- VLAC (Zhai et al., 2025) learns language-conditioned progress differences and completion judgments using trajectory-time supervision -- can be a precedent for learned progress assessment.
- Critic in the Loop (Yi et al., 2026): an independent visual critic predicts discretized, trajectory-derived progress and controls transitions between VLA execution and high-level subtask generation.
- What Matters in Orchestrating Robot Policies (Hu et al., 2026) formalizes a common hierarchical control loop and systematically studies planners, controllers, termination conditions, observation representations, and memory.
The integration and empirical results are promising, but the broad agent loop alone does not establish novelty.
- Some comparisons do not isolate the coordination benefit
- Removing orchestration tests the value of hierarchy, rather than superiority over an existing hierarchy.
- Orchestra with the base policy achieves 2.884 ASL versus HarnessVLA's 2.986; the full system's 7.350 also includes policy adaptations.
Questions for authors
- How many attempts were excluded for each configuration, and were interruptions external or policy-induced?
- Does the improvement persist when policy-related interruptions count as failures?
- What distinguishes Orchestra's progress evaluator and coordination mechanism from existing work, and what evidence shows the benefit of those differences?