Cyborg Index · July 2026

Cirrus-05 vs Vector-04

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

Cirrus-05

Vendor Echo

2.71of 5

Rank

Coverage

33/40

Proof

2.03

Rank down 3 to 7 (was 4)

Vector-04

Vendor Delta

3.07of 5

Rank

Coverage

31/40

Proof

1.75

Rank down 1 to 4 (was 3)

The short answer

Vector-04 scores 0.36 higher than Cirrus-05, and independent evidence backs it.

57% of Vector-04's score is carried by independently verified material rather than claims alone. Cirrus-05 has more unscored ground than the gap between them, so this could change next month.

Where Cirrus-05 is ahead

Deployment+0.77Independently backed
HRI+0.63Independently backed

Where Vector-04 is ahead

Manipulation+1.41Independently backed
Learning+0.85Independently backed
Autonomy+0.80Rests on claims and demos
Perception+0.59Independently backed

What would change it

6 sub-axes on the trailing side have no public evidence yet. If all of them scored full marks they would be worth 0.75 — against a gap of 0.36.

Perception · Sensor stackup to +0.125
Perception · Spatial memoryup to +0.125
Locomotion · Stairsup to +0.125
Locomotion · Uneven terrainup to +0.125
Safety · E stop archup to +0.125

There is more unscored ground than there is gap, so this result could turn over after the next research month.

How much we know

80%of sub-axes covered

We have public evidence for 80% of the 40 sub-axes across these two.

Cirrus-05 · 83%

Vector-04 · 78%

The shape of each score

Cirrus-052.71

33/40 sub-axes have evidence

Faint outline = Vector-04

Vector-043.07

31/40 sub-axes have evidence

Faint outline = Cirrus-05

Reading the wheels — clockwise from the top

1 Locomotion
2 Manipulation
3 Perception
4 Autonomy
5 Learning
6 HRI
7 Deployment
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

Manipulation1.41 Learning0.85 Autonomy0.80 Deployment0.77 HRI0.63 Perception0.59 Locomotion0.36 Safety0.30

Cirrus-05Vector-04

Shared 0–5 track · widest gap first

Manipulation

2.63 / 4.03► 1.41

Independently backedEvidenced cells · 5/5 vs 4/5

Grasp variety

2.65 / 4.11► 1.46

Bimanual

2.58 / 4.07► 1.49

Force control

2.96 / 3.72► 0.76

In hand reorient

2.37 / 4.35► 1.98

Tool use

2.57 / 3.92► 1.35

Learning

2.63 / 3.48► 0.85

Independently backedEvidenced cells · 5/5 vs 3/5

Imitation learning

3.07 / 3.73► 0.66

Sim to real

2.27 / 3.15► 0.88

Foundation policy

2.13 / 3.71► 1.58

Data flywheel

3.06 / 3.15► 0.09

Transfer fine tune

2.62 / 3.64► 1.02

Autonomy

2.51 / 3.31► 0.80

Rests on claims and demosEvidenced cells · 5/5 vs 3/5

Unattended operation

2.53 / 3.57► 1.04

Task horizon

2.11 / 3.63► 1.52

Generalization

2.79 / 3.22► 0.43

Instruction following

2.56 / 3.27► 0.71

Recovery from failure

2.54 / 2.85► 0.31

Deployment

2.50 / 1.73◄ 0.77

Independently backedEvidenced cells · 4/5 vs 4/5

Named pilots

1.97 / 2.10► 0.13

Field hours

2.06 / 1.52◄ 0.54

Customer diversity

2.87 / 1.79◄ 1.08

Supportability

2.87 / 1.31◄ 1.56

Commercial posture

2.73 / 1.93◄ 0.80

HRI

3.20 / 2.57◄ 0.63

Independently backedEvidenced cells · 5/5 vs 5/5

Speech io

3.38 / 2.22◄ 1.16

Nl instruction

3.41 / 2.85◄ 0.56

Intent recognition

3.14 / 2.96◄ 0.18

Safe proximity

3.39 / 2.14◄ 1.25

Social presentation

2.68 / 2.67

Perception

2.90 / 3.49► 0.59

Independently backedEvidenced cells · 2/5 vs 4/5

Sensor stack

2.83 / 3.23► 0.40

Scene understanding

3.37 / 3.97► 0.60

Spatial memory

3.09 / 3.69► 0.60

Object permanence

2.44 / 3.25► 0.81

Semantic world model

2.78 / 3.33► 0.55

Locomotion

2.96 / 3.32► 0.36

Independently backedEvidenced cells · 3/5 vs 4/5

Flat ground walking

3.06 / 3.05

Uneven terrain

2.86 / 3.38► 0.52

Stairs

3.02 / 3.03

Disturbance recovery

2.96 / 3.57► 0.61

Speed endurance

2.90 / 3.59► 0.69

Safety

2.36 / 2.66► 0.30

Independently backedEvidenced cells · 4/5 vs 4/5

Functional safety

2.75 / 3.13► 0.38

Certifications

2.36 / 3.22► 0.86

E stop arch

2.03 / 2.26► 0.23

Incident transparency

2.47 / 2.33◄ 0.14

Force impact limits

2.21 / 2.37► 0.16

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.

Level

Cirrus-05 and Vector-04 read the same on Deployment and Manipulation (2.54 vs 2.50).

Running without a human in the loop

Unattended operation and task horizon, backed by what the robot can perceive.

Vector-04

Vector-04 scores higher on Autonomy and Perception — 3.37 against 2.64 for Cirrus-05.

Moving through difficult ground

Uneven terrain, stairs and disturbance recovery.

Vector-04

Vector-04 scores higher on Locomotion — 3.32 against 2.96 for Cirrus-05.

Working alongside people

Functional safety and certification, plus how it reads and answers people.

Level

Cirrus-05 and Vector-04 read the same on Safety and HRI (2.64 vs 2.63).

Picking up new tasks

Imitation learning, transfer and generalisation across tasks.

Vector-04

Vector-04 scores higher on Learning and Autonomy — 3.42 against 2.59 for Cirrus-05.

Research and developer use

How much of the stack is documented and reproducible.

Vector-04

Vector-04 scores higher on Learning and Perception — 3.48 against 2.72 for Cirrus-05.

How solid the record is on each side

Cirrus-05

Sub-axes on the record: 33 of 40 · 83%
Carried by verified material: 75%
Rests on claims and demos: 0.68

Best documented — Locomotion, Manipulation, Autonomy, Learning, HRI, Deployment, Safety

Nothing on the record — every axis has at least one read

Vector-04

Sub-axes on the record: 31 of 40 · 78%
Carried by verified material: 57%
Rests on claims and demos: 1.32

Best documented — Locomotion, Manipulation, Perception, Autonomy, Learning, HRI, Deployment, Safety

Nothing on the record — every axis has at least one read

What this comparison cannot settle

Neither humanoid has a public read on 1 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.
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