As of: 2026-11-13 13:49 CET.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/
Canonical page: https://postcutoff.com/m/pi-0-5/

# π0.5

Physical Intelligence, π (pi) family. Status: current. Type: robotics model. Released: 2025-04-22.

## How to call it

| Provider | Model id | Endpoint | Docs |
|---|---|---|---|
| GitHub (openpi, JAX + PyTorch) | `gs://openpi-assets/checkpoints/pi05_base` |  | https://github.com/Physical-Intelligence/openpi |
| Hugging Face (LeRobot port) | `lerobot/pi05_base` |  | https://huggingface.co/docs/lerobot/en/pi05 https://huggingface.co/lerobot/pi05_base |

## Pricing

- Price not published.
- Last checked: 2026-09-29

## Specs

- Input: text, image
- Output: action, text
- Context window: not published
- Max output: not published
- Open weights: yes
- Licence: apache-2.0 (openpi code); weights built on PaliGemma, Gemma terms apply

## Where it stands

- Newer robotics models in the registry since its release: 26. The 5 newest:
  - Praxis-1 (2026-09-30) https://postcutoff.com/m/praxis-1/
  - Helix 2.5 (2026-09-17) https://postcutoff.com/m/figure-helix-2-5/
  - UnifoLM-WLA-1.0 (2026-09-10) https://postcutoff.com/m/unifolm-wla-1-0/
  - Skild S1 (Skild Brain) (2026-08-25) https://postcutoff.com/m/skild-s1/
  - Generalist GEN-1.5 (2026-08-19) https://postcutoff.com/m/generalist-gen-1-5/
- Before it from Physical Intelligence: π0-FAST (2025-01-16) https://postcutoff.com/m/pi-0-fast/
- After it from Physical Intelligence: π0.6 / π*0.6 (2025-11-17) https://postcutoff.com/m/pi-0-6/
- Weights are public. Licence: apache-2.0; weights built on PaliGemma, Gemma terms apply. See the licence for commercial terms.

## What stands out

- Open-world generalization to unseen homes: Cleans kitchens and bedrooms in entirely new homes not in training; performance improved as training grew from 3 to 104 homes; co-trained on heterogeneous robot, web and verbal-instruction data. Source: https://www.pi.website/blog/pi05
- Hierarchical subtask prediction + actions in one model: Predicts a high-level text subtask, then low-level actions; trained with knowledge insulation. Source: https://github.com/Physical-Intelligence/openpi
- Newest open-weights π model (found after launch): Base plus LIBERO and DROID checkpoints released in openpi in September 2025; the latest PI model with public weights as of 2026-09 (π0.6/π0.7 are closed). Source: https://github.com/Physical-Intelligence/openpi

## Notes

Announced 2025-04-22; weights open-sourced Sept 2025 (pi05_base, pi05_libero, pi05_droid). Still the most capable open-weights PI model. Full fine-tuning needs >70 GB VRAM.

Sources: [π0.5 blog](https://www.pi.website/blog/pi05), [openpi](https://github.com/Physical-Intelligence/openpi), [LeRobot docs](https://huggingface.co/docs/lerobot/en/pi05).

Model page: https://postcutoff.com/m/pi-0-5/