π₀ → π₀.5: A Walked Tour
Six months. Same architecture. New training recipe. A policy that sometimes works in a kitchen it has never seen.
What actually changed between π₀ (Oct 2024) and π₀.5 (Apr 2025)? Almost nothing in the architecture. Almost everything in the training recipe. This site is a 25-minute walk through the five load-bearing ideas that make that sentence honest — starting at multi-head attention and ending at the open-world generalization claim, framed the way PI's own ablations frame it.
The detour through FAST and Knowledge Insulation is mandatory. FAST is the tokenizer that nearly became π₀-FAST — an autoregressive runtime sister of π₀ that didn't survive deployment — whose tokens now live on as a training-time auxiliary signal. Knowledge Insulation is how those tokens coexist with the flow-matching action head without corrupting the VLM. Together they are the recipe; π₀.5 is the result.
Each page anchors one mental model and one or two visualizations that earn their pixels. Stop wherever you have what you came for — every footer calls out the standalone takeaway. About five minutes a page; you can read the climax (Page 4) on its own if that is all you came for.
Sources: π₀ (arXiv 2410.24164), FAST (arXiv 2501.09747), π₀.5 (arXiv 2504.16054), Knowledge Insulation (arXiv 2505.23705), Real-Time Chunking (arXiv 2506.07339).
How to read this site
Two reasonable paths.
Linear (first read)
- 1 → 2 → 3 → 4 → 5. The arc lands the “aha” on Page 4 (the gradient stops here) and the commercial-relevance click on Page 5 (therefore, novel houses).
- Don't skip Page 1 even if it feels slow. Pages 4 and 5 reuse the prefix-LM mask vocabulary built on Page 1.
Reference (second read / lookup)
- What's an action chunk? → Page 2.
- What is FAST? → Page 3 (and the linked FAST Workshop for the tokenizer pipeline in detail).
- Why doesn't π₀.5 just freeze the VLM? → Page 4 — this is the most-misunderstood detail on the whole site.
- What concretely changed between π₀ and π₀.5? → Page 5.