THE RESEARCH GUIDE
Before you compare,
ask better questions.
A practical reading checklist for humanoid specifications and demonstrations. Use it to make the limits of the evidence explicit.
Pin down the configuration.
Write down the manufacturer, product, generation and hardware options. Do the photograph, demonstration and specification sheet describe the same configuration?
A G1 EDU simulation model does not establish standard G1 hand specifications.
Separate the claim from the test.
Who made the statement? Is it a manufacturer claim, reporting that repeats that claim, or a documented independent test? Keep the original source and its date.
Roboryn’s editorial approval means a claim was selected for display; it does not mean the robot was tested.
Keep units and conditions attached.
For runtime, ask about task, payload and duty cycle. For load, ask about reach and movement. For speed, ask about terrain and measurement method.
An approximate runtime without a stated workload cannot predict runtime in your own application.
Treat relative improvements carefully.
Ask for the baseline, measurement method and absolute result. Without these, a percentage or reduction describes a comparison whose limits still need explaining.
A claimed mass reduction is not total mass. A percentage speed increase is not walking speed in metres per second.
Interrogate the demonstration.
Ask whether motion was autonomous, teleoperated or scripted; how many attempts were made; what was edited; and what support or environment preparation was required.
A successful clip alone does not establish repeatability, general task autonomy or readiness for deployment.
Check the decision date.
Record when a source was published, when it was captured and when it was reviewed. Check availability and commercial terms directly before making a purchase decision.
A historical starting price is not a current quote or a promise of delivery.
Try it on the current evidence.
Compare scoped body heights and joint degrees of freedom across G1, H1 and T1, while keeping Optimus Gen 2’s relative claims separate.
Open the comparisonA FREE RESEARCH SAMPLE
Take the evidence with you.
Four configurations. Which comparisons hold? Sample 02 explores four research questions using nine selected claims. The companion CSV contains all 17 approved claims, with definitions, conditions, original expressions, review type and source dates.
Download the evidence brief (PDF) ↓
Download all 17 approved claims (CSV) ↓
Based on the edition captured 2026-10-18. G1 and Optimus retain their September evidence; H1 and T1 were researched in October. This is not a simultaneous refresh of all four records. All statements remain manufacturer claims without independent verification. The G1 price is historical. No account or email required.
How to read and reuse the CSV
One row is one selected claim for an exact generation. Keep the definition, units, qualifier and conditions together. Blank numeric cells are not zero; a relative improvement is not an absolute measurement. The file includes original source URLs and locations, capture dates, editorial decision dates and actual source-review type.
Original source text and third-party models retain their own rights. This sample does not grant rights to redistribute those assets. Read the field guide.
Help make this reference useful.
Which research question brought you here? Tell us what was useful or missing. Please include a source if you suggest a new claim.
Share research feedbackOpens your email app. No newsletter signup, account or automatic submission. How contact data is handled.