MAGMA-Bench: Evaluating Long-Horizon Embodied Agents under Dynamic Tasks and Execution Failures
We introduce MAGMA-Bench, an architecture-agnostic benchmark for persistent task reasoning in dynamic multi-robot environments.
ProofPaper ↗
Key points
- Long-horizon embodied agents must maintain task-relevant information, coordinate heterogeneous robots, and adapt their decisions as execution unfolds.
- Existing benchmarks investigate dynamic task execution and embodied multi-agent coordination, but provide limited evaluation of their intersection.
- MAGMA-Bench evaluates agents across increasing task horizons and controlled online perturbations, including execution failures, task interruptions, and mission updates, with scenarios requiring coordination across robots with heterogeneous capabilities.
- We release the benchmark, 261,646 task-level training examples, and an architecture-agnostic expert-trace generation pipeline.
Sources (1)
- [1]MAGMA-Bench: Evaluating Long-Horizon Embodied Agents under Dynamic Tasks and Execution FailuresarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 9, 12:54 PM
We introduce MAGMA-Bench, an architecture-agnostic benchmark for persistent task reasoning in dynamic multi-robot environments.
Long-horizon embodied agents must maintain task-relevant information, coordinate heterogeneous robots, and adapt their decisions as execution unfolds.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 9, 2026Beyond Successful Robot Demonstrations: Learning from Human Interactions and Modeling Failure Dynamics
- Oct 9, 2026Minions: A Dynamic and Distributed Virtual GPU Cloud Built on Idle Apple Silicon Devices
- Oct 9, 2026Rethinking Relative Policy Differences Transfer of On-Policy Distillation
- Oct 9, 2026InstantFusion: A Shared Latent Interface for Heterogeneous Image Generators
- Oct 9, 2026Towards Physically Grounded Articulated 3D Assets: A Multi-Agent Pipeline and Benchmark
