METHOD & ACCOUNTABILITY
Evidence first.
Limits included.
Roboryn is an independent research project by Pascal Hesselmeier. It is not affiliated with Unitree, Tesla or Booster, and does not independently test robots.
What this edition contains
Seventeen selected claims: five for standard Unitree G1, eight for Tesla Optimus Gen 2, and two each for original basic H1 and Booster T1 with 23 DoFs. These are manufacturer claims, reviewed for inclusion. This edition was exported on 2026-10-18; it is not a live view of manufacturer updates.
The G1 evidence was captured on 6 September 2026 and approved on 12 September. The Optimus evidence was captured and approved on 13 September; its publication was recorded on 14 September. H1 and T1 evidence was captured on 18 October 2026. Their source review was performed by agents with an independent agent challenge, and the owner authorized the four exact qualified claims. That record does not claim personal source inspection by the owner. A publication or website update does not reset earlier dates.
Four separate questions
- Identity
- Which robot generation and configuration does the evidence describe?
- Provenance
- Who originally made the claim? Repeated reporting of one claim is not independent corroboration.
- Verification
- Has the claim been independently checked under documented conditions? All claims in this edition remain not independently verified.
- Editorial selection
- Which retained claims were approved for this record? Inclusion does not prove real-world performance.
Unknown is not zero
We retain original values, units, configuration, source attribution and known dates. Approximate values stay approximate. Relative improvements stay relative. Missing evidence remains explicitly unknown.
Three configurations share scoped height and whole-body DoF fields. H1 height remains approximate; model and hardware options stay explicit. The comparison does not establish a more capable robot, a precise clearance requirement or a speed ranking.
Sources and history
Profile pages link to the original source and show the retained value or precise locator. Original pages can change. Corrections and later source captures must preserve earlier attribution and meaning; they receive a separate, attributable editorial decision before publication. Routine source review and publication are delegated, with an independent agent challenge of the underlying evidence. Research imports alone cannot publish, and automated review is never labelled as human inspection or an independent robot test.
We aim to revisit prominent price and availability claims within 30 days and stable physical specifications within 180 days. These are editorial targets, not automatic refresh promises. The G1 price is outside that review window and is shown as historical evidence only.
How to read the 3D models
The G1 uses an upstream simulation asset whose configuration differs in some details from the standard specification record. The Optimus viewer is an original, approximate visual reconstruction, not Tesla CAD or a dimensionally calibrated model.
Assembly views show available mesh groups. Motion studies are illustrative animation, not evidence of autonomy or task capability. Each viewer includes attribution and model notes. H1 and T1 currently have sourced text profiles without a 3D model.
CORRECTIONS & FEEDBACK
Help improve the record.
Email hallo@cashflowetf.de with “Roboryn” in the subject. Include the profile, generation/configuration, original source link, relevant passage and what needs correcting.
Submissions are research leads, not automatic edits. Accepted changes need an attributable editorial decision and a new published edition.
Send a source correctionOpens your email app. Privacy information.
Independence
This edition is free and carries no advertising, affiliate links, sponsorships or paid rankings. Commercial relationships must never decide which robot claims are selected.
Operator information / Impressum