---
title: RoboCerebra
description: "NeurIPS 2025 long-horizon household bench (1,000 trajectories, 100 variants). WorldFoundry imports official rollouts; it does not run LIBERO."
---

RoboCerebra (Beihang CoLa Lab, NeurIPS 2025) tests **System-2** planning on long household sequences, not reactive System-1 control. The release is **1,000** human-annotated trajectories over **100** task variants (mean **9.1** atomic steps; horizons on the order of **3,000** simulation steps) on a **LIBERO**-based stack. Published eval is **600** rollouts (**60** tasks × **10** trials) with anchor-aligned planner / VLA handoffs. Six official conditions: **Ideal**, **Memory Exploration**, **Memory Execution**, **Random Disturbance**, **Observation Mismatching**, and **Mix**.

Paper: [RoboCerebra](https://arxiv.org/abs/2506.06677). Project: [robocerebra.github.io](https://robocerebra.github.io/). Code: [buaa-colalab/RoboCerebra](https://github.com/buaa-colalab/RoboCerebra).

WorldFoundry **imports** official result exports (`worldfoundry-eval zoo benchmark-run --mode official-validation`). It does **not** run the bundled LIBERO simulator, collect rollouts, or host the hierarchical VLM planner + VLA controller. Declare the six-way condition split on the export if you want comparable numbers.

<BenchmarkPageSections id="robocerebra" locale="en" />

## How to run

```bash
worldfoundry-eval zoo benchmark-run \
  --benchmark-id robocerebra \
  --mode official-validation \
  --official-results-path <official_results.json> \
  --output-dir tmp/robocerebra/official-validation \
  --json
```

Optional `WORLDFOUNDRY_ROBOCEREBRA_ROOT` points at a local bench/runtime root. Default import path: `WORLDFOUNDRY_ROBOCEREBRA_RESULTS_PATH`. Manifest: `worldfoundry/data/benchmarks/tasks/external/robocerebra.yaml`. Outputs: `scorecard.json` (primary `success_rate`), `raw_results.jsonl`, `evaluation/summary.json`.

[← Benchmark Hub](/docs/evaluation/benchmark-hub)