LeRobot v0.6.0
Hugging Face's open-source robotics toolkit adds world-model policies, reward models, and simulation benchmarks to close the robot learning loop.
Overview
Hugging Face released LeRobot v0.6.0 on July 7, 2026, extending its open-source embodied-AI toolkit from data collection into a full imagine-evaluate-improve loop. The release adds world-model policies (VLA-JEPA, FastWAM, LingBot-VA), more vision-language-action models, reward-model support, six simulation benchmarks under lerobot-eval, a rollout CLI that turns failures into training data, FSDP training, and HF Jobs cloud training, plus depth sensing and VLM-powered dataset annotation. It also ships breaking changes: `pip install lerobot` no longer bundles dataset/training dependencies by default, and some internal import paths were replaced with canonical public entry points.
This is a sourced directory entry. Features and capabilities are reported in the linked sources; this entry does not document independent Aymorphic testing.
Key features
- World-model policies: VLA-JEPA, FastWAM, LingBot-VA
- Reward-model support and six simulation benchmarks via lerobot-eval
- Rollout tooling that converts human corrections into training data
- FSDP training and HF Jobs cloud training
- Depth sensing and VLM-powered dataset annotation; leaner default install
Use cases
- Training and evaluating vision-language-action robot policies end to end
- Benchmarking robot learning policies in simulation before real-world deployment
- Cloud-based robot policy training via HF Jobs
Pricing
Free and open source (Apache 2.0-style Hugging Face OSS license); HF Jobs cloud training billed per Hugging Face's standard compute pricing.
Pricing reflects the source record and may change. Confirm current terms with the provider.
Dates and sources
- Released
- 7 Jul 2026
- Record updated
- 23 Sept 2026
- Record last checked
- 23 Sept 2026