CYBORGS OF EARTH
Cyborg Index · July 2026

Ember-09 vs Onyx-08

Both humanoids read on the same 8 axes and 40 sub-axes, using the published Index scores. Every difference below carries the public evidence behind it, and sub-axes with no public evidence count as 0 rather than being skipped.

Overall standing

Capability score · 0–5 over 40 sub-axes
Ember-09
Vendor India
1.90of 5
Rank
#11
Coverage
29/40
Proof
1.03
Rank down 3 to 11 (was 8)
Onyx-08
Vendor Hotel
2.74of 5
Rank
#6
Coverage
25/40
Proof
1.20
The short answer

Onyx-08 scores 0.84 higher than Ember-09, but that lead rests mostly on vendor claims and demos.

Only 44% of Onyx-08's score is carried by independently verified material — the rest is claimed or shown in demos. Ember-09 cannot close the gap on unscored ground alone.

Where Ember-09 is ahead
  • Deployment+0.46Rests on claims and demos
Where Onyx-08 is ahead
  • Manipulation+1.57Rests on claims and demos
  • Autonomy+1.35Rests on claims and demos
  • Perception+1.14Rests on claims and demos
  • Locomotion+1.03Rests on claims and demos
What would change it

5 sub-axes on the trailing side have no public evidence yet. If all of them scored full marks they would be worth 0.63 — against a gap of 0.84.

  • HRI · Social presentationup to +0.125
  • HRI · Speech ioup to +0.125
  • Autonomy · Task horizonup to +0.125
  • Deployment · Named pilotsup to +0.125
  • Deployment · Customer diversityup to +0.125
How much we know
68%of sub-axes covered

We have public evidence for 68% of the 40 sub-axes across these two.

Ember-09 · 73%
Onyx-08 · 63%
The shape of each score
Ember-091.90
LocomotionManipulationPerceptionAutonomyLearningHRIDeploymentSafety

29/40 sub-axes have evidence

Faint outline = Onyx-08

Onyx-082.74
LocomotionManipulationPerceptionAutonomyLearningHRIDeploymentSafety

25/40 sub-axes have evidence

Faint outline = Ember-09

Reading the wheels — clockwise from the top
  1. 1 Locomotion
  2. 2 Manipulation
  3. 3 Perception
  4. 4 Autonomy
  5. 5 Learning
  6. 6 HRI
  7. 7 Deployment
  8. 8 Safety

Each wheel has 40 spokes — five per axis. A longer spoke is a higher score, and the dashed rings mark 1 to 4. Hover or tab onto any spoke to read the axis, the sub-axis and the exact value. A missing spoke means there is no public evidence for that sub-axis yet, which is not the same as scoring zero. The faint outline on each wheel is the other humanoid, drawn on the same scale.

Every axis, every sub-axis
Manipulation1.57Autonomy1.35Perception1.14Locomotion1.03Learning0.95Safety0.66HRI0.47Deployment0.46
Ember-09Onyx-08
Shared 0–5 track · widest gap first
Manipulation
2.00 / 3.57► 1.57
Rests on claims and demosEvidenced cells · 5/5 vs 5/5
Grasp variety
1.95 / 3.71► 1.76
Bimanual
1.95 / 3.12► 1.17
Force control
2.30 / 3.77► 1.47
In hand reorient
1.58 / 3.53► 1.95
Tool use
2.24 / 3.73► 1.49
Autonomy
1.71 / 3.06► 1.35
Rests on claims and demosEvidenced cells · 4/5 vs 4/5
Unattended operation
2.19 / 2.59► 0.40
Task horizon
2.03 / 3.28► 1.25
Generalization
1.52 / 3.44► 1.92
Instruction following
1.50 / 2.72► 1.22
Recovery from failure
1.31 / 3.27► 1.96
Perception
2.16 / 3.30► 1.14
Rests on claims and demosEvidenced cells · 4/5 vs 3/5
Sensor stack
1.99 / 3.62► 1.63
Scene understanding
2.11 / 3.25► 1.14
Spatial memory
2.30 / 2.95► 0.65
Object permanence
2.01 / 3.41► 1.40
Semantic world model
2.38 / 3.25► 0.87
Locomotion
2.30 / 3.33► 1.03
Rests on claims and demosEvidenced cells · 5/5 vs 5/5
Flat ground walking
2.15 / 3.01► 0.86
Uneven terrain
2.62 / 3.74► 1.12
Stairs
2.53 / 3.54► 1.01
Disturbance recovery
2.16 / 3.46► 1.30
Speed endurance
2.06 / 2.91► 0.85
Learning
1.87 / 2.83► 0.95
Not enough public record yetEvidenced cells · 0/5 vs 0/5
Imitation learning
1.66 / 3.07► 1.41
Sim to real
1.93 / 2.90► 0.97
Foundation policy
2.22 / 2.76► 0.54
Data flywheel
1.76 / 2.99► 1.23
Transfer fine tune
1.80 / 2.42► 0.62
Safety
1.87 / 2.53► 0.66
Rests on claims and demosEvidenced cells · 5/5 vs 0/5
Functional safety
1.87 / 2.50► 0.63
Certifications
2.02 / 2.45► 0.43
E stop arch
1.93 / 2.23► 0.30
Incident transparency
1.96 / 2.62► 0.66
Force impact limits
1.56 / 2.86► 1.30
HRI
1.86 / 2.33► 0.47
Rests on claims and demosEvidenced cells · 3/5 vs 4/5
Speech io
2.15 / 2.42► 0.27
Nl instruction
1.83 / 2.45► 0.62
Intent recognition
1.76 / 2.25► 0.49
Safe proximity
1.99 / 1.84◄ 0.15
Social presentation
1.59 / 2.70► 1.11
Deployment
1.38 / 0.93◄ 0.46
Rests on claims and demosEvidenced cells · 3/5 vs 4/5
Named pilots
1.64 / 1.02◄ 0.62
Field hours
0.95 / 1.06► 0.11
Customer diversity
1.16 / 0.73◄ 0.43
Supportability
1.46 / 0.88◄ 0.58
Commercial posture
1.71 / 0.95◄ 0.76
Segment length and thickness are the gapNo public evidence yet
Which is better for what

Each row is decided by the axes that matter for that kind of work, read off the same Capability scores as the ranking. Rows only appear where both humanoids have enough public record to answer.

Manufacturing and warehouse work
Named pilots and field hours, backed by what the hands can actually do.
Onyx-08
Onyx-08 scores higher on Deployment and Manipulation — 1.81 against 1.59 for Ember-09.
Running without a human in the loop
Unattended operation and task horizon, backed by what the robot can perceive.
Onyx-08
Onyx-08 scores higher on Autonomy and Perception — 3.14 against 1.86 for Ember-09.
Moving through difficult ground
Uneven terrain, stairs and disturbance recovery.
Onyx-08
Onyx-08 scores higher on Locomotion — 3.33 against 2.30 for Ember-09.
Working alongside people
Functional safety and certification, plus how it reads and answers people.
Onyx-08
Onyx-08 scores higher on Safety and HRI — 2.47 against 1.87 for Ember-09.
Picking up new tasks
Imitation learning, transfer and generalisation across tasks.
Onyx-08
Onyx-08 scores higher on Learning and Autonomy — 2.91 against 1.82 for Ember-09.
Research and developer use
How much of the stack is documented and reproducible.
Onyx-08
Onyx-08 scores higher on Learning and Perception — 2.98 against 1.97 for Ember-09.
How solid the record is on each side
Ember-09
Sub-axes on the record
29 of 40 · 73%
Carried by verified material
54%
Rests on claims and demos
0.87
Best documented — Locomotion, Manipulation, Perception, Autonomy, HRI, Deployment, Safety
Nothing on the record — Learning
Onyx-08
Sub-axes on the record
25 of 40 · 63%
Carried by verified material
44%
Rests on claims and demos
1.54
Best documented — Locomotion, Manipulation, Perception, Autonomy, HRI, Deployment
Nothing on the record — Learning, Safety

What this comparison cannot settle

  • Neither humanoid has a public read on 6 of the 40 sub-axes, so those cannot be compared at all.
  • 14 sub-axes are read on only one side. A one-sided cell is undecided, not a zero for the silent side.
  • No public evidence on either side for Learning.
  • Demonstrations happen in different environments and product revisions are not always equivalent, so like-for-like uptime and throughput figures do not exist for most of the field.

Go deeper on either humanoid

Nearby comparisons

All comparisons →
Cyborg IndexJuly 2026 editionCapability scoremethodologypreview cohort