RoboMME adds a 16-task benchmark for robot long-horizon memory
chris_j_paxton · x · 2026-07-25
RoboMME is presented as a new benchmark for robot long-horizon memory, aimed at measuring whether generalist policies can understand language and carry out multi-step tasks over time.
- It spans 16 robot tasks, including counting objects and timing-sensitive manipulation.
- The authors evaluate 14 memory-augmented generalist policies across those tasks.
- The paper argues that robotics still lacks a strong benchmark for memory, even though memory is central to manipulation and instruction following.
- The work is framed as a step toward making progress on a core capability rather than just ad hoc demos.
Related event: RoboMME: A New Benchmark for Evaluating Robot Long-term Memory(4 posts)→
More from Embodied
- Hypothesis-Driven Robots Reach 93% Success in Real-World Tasks Using GPT-4.1 — imjustnewatai · 2026-07-25
- MIT project on multi-agent inverse design wins a Genesis Mission award — ProfBuehlerMIT · 2026-07-25
- Mondo’s Beni pairs cute hardware with millions of simulated RL trials — AnandSwa · 2026-07-25
- Almond Robotics Sells Out First Batch, Starts Production on V2 — chris_j_paxton · 2026-07-25
- Robotics research is mostly limited by robot hours or GPU hours, says one researcher — chris_j_paxton · 2026-07-25
- A cotton-picking robot with 108 arms is the latest robotics spectacle — chris_j_paxton · 2026-07-25