Skip to content
Bud's Wiki

Eren Chen teases Physical Intelligence π0.8 announcement

TL;DR

Section titled “TL;DR”

A two-word teaser from Physical Intelligence researcher Eren Chen (“Pi 0.8 announcement 👀”) with a single screenshot image, posted Sep 30 2026. First public signal of a π0.8 successor in the Physical Intelligence VLA lineage (π0 → π0.5 → π0.6 → π0.7 → π0.8). No linked blog, report, or weights at filing time; attached image not readable at ingest.

Key claims

Section titled “Key claims”
  • Chen is signaling that π0.8 is imminent or just-shipped from Physical Intelligence, visible in a screenshot embedded in the tweet [tweet body + attached image].

Method

Section titled “Method”

Not applicable — standalone tweet with an unlabeled image attachment. Image contents could not be transcribed at filing time.

Results

Section titled “Results”

None reported. 5,800 views, 85 likes, 21 reposts as of fetch time.

Why it’s interesting

Section titled “Why it’s interesting”

Signals the π line is still on its roughly-quarterly cadence through the back half of 2026 (π0.7 landed in April 2026 per π0.7: A Steerable Robotic Foundation Model with Emergent Compositional Generalization, π0.6 / RECAP is already filed as π*0.6: a VLA That Learns From Experience (RECAP)). Worth a /bud refresh once Physical Intelligence publishes the companion blog or technical report so this page can be reconciled with the concrete π0.8 recipe deltas — current filed π-line entries show the lineage drifting from action-pretraining (π0.5) through RL-via-advantage-conditioning (π0.6 / RECAP) to steerable composition (π0.7), and the key open question on VLA Models is where π0.8 lands on those axes.

See also

Section titled “See also”