Find my robot
Illustrative blueprint humanoid carrying dishes with task-evidence graphics
Home›Intelligence›Humanoid capability tracker
RightRobot living tracker

Humanoid capability tracker: what can each robot actually do?

A single task-by-task view of NEO, Figure 03, Unitree G1 and Optimus — designed to be updated as the evidence changes instead of forcing readers to piece together manufacturer demos and headlines themselves.

Illustrative RightRobot concept artwork
RightRobot home robotics

Move between the market, the evidence and what you can buy

SignalsRightRobot Intelligence

Important launches, records, evidence shifts and market reality checks.

Open Intelligence →Future-facingHumanoid robots

NEO, Figure, Unitree, Optimus and the home-humanoid evidence layer.

Explore humanoids →BridgeFuture Tech

Stair-climbing vacuums, robotic arms and specialist robots becoming more capable.

See Future Tech →Useful nowFind my robot vacuum

Turn home-robot research into a real current buying decision.

Run Home Fit →

Capability matrix

This is deliberately conservative. “Demonstrated” does not silently become “reliable autonomous consumer feature”.

Capability1X NEOFigure 03Unitree G1Tesla Optimus
Consumer order pathOpen in USNo public checkoutOfficially orderable; currently backorderedNo public checkout
Designed for home choresYesYesNot turnkey-home focusedFuture ambition
Dishwasher / dishesAutonomous demo documentedCompany says dishes autonomouslyNot established as turnkey home taskNot enough current consumer evidence
LaundryMarketed; Expert Mode may assistCompany says laundry autonomouslyNot established as turnkey home taskNot enough current consumer evidence
StairsManufacturer-documented AI mobilityFigure says home stair navigationStrong mobility platform; home task relevance unclearNot enough current home evidence
CookingTeleoperated showcase onlyNot establishedNot establishedNot established
Long unsupervised multi-chore housekeepingNot independently provenNot independently provenNot a turnkey-home propositionNot independently proven
Independent consumer-home evidenceVery limitedVery limitedPlatform evidence stronger than home evidenceVery limited
1X NEO
Consumer order path
Open in US
Designed for home chores
Yes
Dishwasher / dishes
Autonomous demo documented
Laundry
Marketed; Expert Mode may assist
Stairs
Manufacturer-documented AI mobility
Cooking
Teleoperated showcase only
Long unsupervised multi-chore housekeeping
Not independently proven
Independent consumer-home evidence
Very limited
Figure 03
Consumer order path
No public checkout
Designed for home chores
Yes
Dishwasher / dishes
Company says dishes autonomously
Laundry
Company says laundry autonomously
Stairs
Figure says home stair navigation
Cooking
Not established
Long unsupervised multi-chore housekeeping
Not independently proven
Independent consumer-home evidence
Very limited
Unitree G1
Consumer order path
Officially orderable; currently backordered
Designed for home chores
Not turnkey-home focused
Dishwasher / dishes
Not established as turnkey home task
Laundry
Not established as turnkey home task
Stairs
Strong mobility platform; home task relevance unclear
Cooking
Not established
Long unsupervised multi-chore housekeeping
Not a turnkey-home proposition
Independent consumer-home evidence
Platform evidence stronger than home evidence
Tesla Optimus
Consumer order path
No public checkout
Designed for home chores
Future ambition
Dishwasher / dishes
Not enough current consumer evidence
Laundry
Not enough current consumer evidence
Stairs
Not enough current home evidence
Cooking
Not established
Long unsupervised multi-chore housekeeping
Not independently proven
Independent consumer-home evidence
Very limited

Why RightRobot does not give humanoids an overall score yet

RightRobot is deliberately withholding a public overall 0–100 robot ranking for now. The evidence is not yet deep or independent enough to justify telling consumers that one early humanoid is numerically “better” than another. The public tracker therefore shows an evidence-depth band and task-level evidence rather than a pseudo-precise confidence percentage.

Current evidence depth: Moderate
1X NEO
Not yet recommendation-ready

Why no public overall score yet: current evidence is too uneven for a precise 0–100 consumer ranking to be responsible. RightRobot therefore shows an evidence-depth band and task-level notes instead.

  • Strongest current consumer-home proposition in this group.
  • Expert Mode and limited independent owner evidence constrain autonomy/intervention scores.
  • Soft/tendon-driven domestic design is a positive safety signal, not a substitute for broad independent home testing.
Current evidence depth: Limited
Figure 03
Not yet recommendation-ready

Why no public overall score yet: current evidence is too uneven for a precise 0–100 consumer ranking to be responsible. RightRobot therefore shows an evidence-depth band and task-level notes instead.

  • Explicitly designed for home tasks and described by Figure as autonomous.
  • No normal consumer price/order path yet, limiting value/support assessment.
  • Independent consumer-home evidence remains limited.
Current evidence depth: Moderate
Unitree G1
Not yet recommendation-ready

Why no public overall score yet: current evidence is too uneven for a precise 0–100 consumer ranking to be responsible. RightRobot therefore shows an evidence-depth band and task-level notes instead.

  • Strong buyable-hardware and mobility benchmark.
  • Not positioned as a turnkey autonomous household helper.
  • Developer/platform orientation suppresses household usefulness and low-intervention scores.
Current evidence depth: Very limited
Tesla Optimus
Not yet recommendation-ready

Why no public overall score yet: current evidence is too uneven for a precise 0–100 consumer ranking to be responsible. RightRobot therefore shows an evidence-depth band and task-level notes instead.

  • High long-term potential but no public consumer purchase route or official retail price.
  • Most household criteria remain future-facing rather than consumer-verifiable.
  • Low evidence confidence is retained internally so the methodology cannot hide weak evidence behind a high raw capability estimate.

Weighting model

Household usefulness
16%
Genuine autonomy
16%
Manipulation & dexterity
12%
Home navigation & mobility
10%
Family safety
10%
Privacy & user control
10%
Low human intervention
8%
Reliability & recovery
8%
Support & repairability
5%
Value
5%

How this becomes Humanoid Home Fit later

When there are enough real consumer choices, the same score can become home-specific. A household with stairs, children, pets, privacy sensitivity or a strong need for laundry/dish handling should receive different recommendations rather than one universal “best humanoid”.