CYBORGS OF EARTH
The Cyborg IndexCapability landscape

Capability landscape · Sep 2026 edition

The whole field on one plot.

  • 01The whole field sits below 1.50 of 5 on capability — 13 ranked machines inside 0.93 of a point.
  • 02Evidence strength averages 10% — 0 of 13 carry a medium or stronger public record.
  • 03Figure 03 moved the most on the record this edition: +0.15 capability.

1. Map view — capability landscape

All 13 offerings positioned by capability (X) against evidence strength (Y).

How to read & controls

How to read

  • X = Capability — what the machine can do on the public record.
  • Y = Evidence strength — how much of that record actually proves it.
  • Each glyph draws the 8 capability axes from its centre.
  • The outer ring reports per-axis evidence.
  • A trail is movement since the previous edition.

View mode

Market lens

This view zooms into the current market.

Capability in viewof 0–5
0.401.67
Evidence in viewof 0–100
226
0.400.721.041.351.672%8%14%20%26%Median of this month’s ranked cohort. Not a threshold.MEDIANMedian of this month’s ranked cohort. Not a threshold.MEDIANCAPABILITY (cohort range of 0–5)EVIDENCE (cohort range of 0–100%)Relative mode magnifies differences inside the current market. Absolute scores do not change.RELATIVE · CURRENT COHORTTesla Optimus: capability moved from 0.73 to 0.78 since the previous editionAgibot A2: capability moved from 1.40 to 1.50 since the previous editionFourier GR-2: capability moved from 0.58 to 0.57 since the previous editionFigure 03: capability moved from 1.20 to 1.35 since the previous edition1X NEO: capability moved from 1.28 to 1.33 since the previous editionApptronik Apollo: capability moved from 0.97 to 0.98 since the previous editionKepler Forerunner K3: capability moved from 1.17 to 1.23 since the previous editionKepler Forerunner K2: capability moved from 1.10 to 1.18 since the previous editionXPeng IRON: capability moved from 0.48 to 0.57 since the previous edition9101Agibot A21.50▲0.10123Figure 031.35▲0.15114872UBTech Walker S1.48±0.005613

Swipe the plot sideways to pan

Evidence strength (all axes)

  • Very weakunder 10%
  • Weak10–30%
  • Medium30–60%
  • Strong60% and above

Movement vs previous edition

  • ↑ Improved 3
  • — Held 7
  • ↓ Declined 3

Inspector

Select any glyph to open its full read-out — eight-axis radar, movement and evidence strength.

This month
Evidence strength averages 10% — 0 of 13 carry a medium or stronger public record.

2. Orbit view — eight-axis orbit

Every offering plotted around the 8 capability axes. Shows who leads each axis.

How to read & controls

How to read

  • Each sector is one capability axis.
  • Distance from the centre is that axis’s published read.
  • A dot is one machine on one axis.
  • Fainter dots are thinly evidenced, not weak.

Leaders

Locomotion

every machine read

Manipulation

every machine read

Perception

every machine read

Autonomy

every machine read

Learning

every machine read

HRI

every machine read

Deployment

every machine read

Safety

every machine read

0.511.52Locomotionevidence 15%Manipulationevidence 8%Perceptionevidence 14%Autonomyevidence 9%Learningevidence 10%HRIevidence 13%Deploymentevidence 2%Safetyevidence 12%Agibot A2 · Locomotion 1.20/5 · evidence 27%UBTech Walker S · Locomotion 1.60/5 · evidence 12%Figure 03 · Locomotion 0.80/5 · evidence 14%1X NEO · Locomotion 1.60/5 · evidence 15%Kepler Forerunner K3 · Locomotion 1.60/5 · evidence 8%Kepler Forerunner K2 · Locomotion 1.60/5 · evidence 8%NEURA 4NE-1 · Locomotion 0.80/5 · evidence 10%Apptronik Apollo · Locomotion 0.80/5 · evidence 11%Tesla Optimus · Locomotion 0.80/5 · evidence 34%Boston Dynamics Atlas (electric) · Locomotion 0.80/5 · evidence 26%Unitree H2 · Locomotion 1.20/5 · evidence 9%Fourier GR-2 · Locomotion 0.80/5 · evidence 14%XPeng IRON · Locomotion 0.40/5 · evidence 8%Agibot A2 · Manipulation 1.20/5 · evidence 11%UBTech Walker S · Manipulation 1.60/5 · evidence 8%Figure 03 · Manipulation 1.40/5 · evidence 5%1X NEO · Manipulation 0.80/5 · evidence 4%Kepler Forerunner K3 · Manipulation 1.20/5 · evidence 8%Kepler Forerunner K2 · Manipulation 1.20/5 · evidence 12%NEURA 4NE-1 · Manipulation 1.20/5 · evidence 4%Apptronik Apollo · Manipulation 1.20/5 · evidence 13%Tesla Optimus · Manipulation 0.80/5 · evidence 8%Boston Dynamics Atlas (electric) · Manipulation 1.20/5 · evidence 12%Unitree H2 · Manipulation 1.00/5 · evidence 6%Fourier GR-2 · Manipulation 0.80/5 · evidence 8%XPeng IRON · Manipulation 0.60/5 · evidence 2%Agibot A2 · Perception 0.80/5 · evidence 25%UBTech Walker S · Perception 1.60/5 · evidence 6%Figure 03 · Perception 1.60/5 · evidence 16%1X NEO · Perception 1.60/5 · evidence 7%Kepler Forerunner K3 · Perception 1.40/5 · evidence 9%Kepler Forerunner K2 · Perception 1.40/5 · evidence 8%NEURA 4NE-1 · Perception 1.00/5 · evidence 8%Apptronik Apollo · Perception 0.60/5 · evidence 12%Tesla Optimus · Perception 0.60/5 · evidence 33%Boston Dynamics Atlas (electric) · Perception 0.40/5 · evidence 33%Unitree H2 · Perception 0.80/5 · evidence 11%Fourier GR-2 · Perception 1.00/5 · evidence 13%XPeng IRON · Perception 1.20/5 · evidence 7%Agibot A2 · Autonomy 1.60/5 · evidence 14%UBTech Walker S · Autonomy 1.80/5 · evidence 3%Figure 03 · Autonomy 1.60/5 · evidence 9%1X NEO · Autonomy 1.00/5 · evidence 4%Kepler Forerunner K3 · Autonomy 1.20/5 · evidence 5%Kepler Forerunner K2 · Autonomy 1.40/5 · evidence 8%NEURA 4NE-1 · Autonomy 1.20/5 · evidence 8%Apptronik Apollo · Autonomy 1.60/5 · evidence 8%Tesla Optimus · Autonomy 1.00/5 · evidence 23%Boston Dynamics Atlas (electric) · Autonomy 0.20/5 · evidence 15%Unitree H2 · Autonomy 0.40/5 · evidence 3%Fourier GR-2 · Autonomy 0.40/5 · evidence 8%XPeng IRON · Autonomy 0.40/5 · evidence 8%Agibot A2 · Learning 2.00/5 · evidence 8%UBTech Walker S · Learning 0.40/5 · evidence 13%Figure 03 · Learning 1.60/5 · evidence 12%1X NEO · Learning 2.00/5 · evidence 8%Kepler Forerunner K3 · Learning 1.60/5 · evidence 12%Kepler Forerunner K2 · Learning 0.80/5 · evidence 6%NEURA 4NE-1 · Learning 1.20/5 · evidence 11%Apptronik Apollo · Learning 0.80/5 · evidence 10%Tesla Optimus · Learning 1.20/5 · evidence 18%Boston Dynamics Atlas (electric) · Learning 0.80/5 · evidence 19%Unitree H2 · Learning 0.20/5 · evidence 5%Fourier GR-2 · Learning 0.20/5 · evidence 5%XPeng IRON · Learning 0.40/5 · evidence 8%Agibot A2 · HRI 2.00/5 · evidence 16%UBTech Walker S · HRI 2.00/5 · evidence 11%Figure 03 · HRI 1.60/5 · evidence 8%1X NEO · HRI 1.60/5 · evidence 11%Kepler Forerunner K3 · HRI 1.60/5 · evidence 5%Kepler Forerunner K2 · HRI 1.60/5 · evidence 5%NEURA 4NE-1 · HRI 1.60/5 · evidence 9%Apptronik Apollo · HRI 1.20/5 · evidence 6%Tesla Optimus · HRI 0.80/5 · evidence 36%Boston Dynamics Atlas (electric) · HRI 0.40/5 · evidence 33%Unitree H2 · HRI 1.00/5 · evidence 14%Fourier GR-2 · HRI 0.40/5 · evidence 13%XPeng IRON · HRI 0.40/5 · evidence 3%Agibot A2 · Deployment 1.60/5 · evidence 4%UBTech Walker S · Deployment 2.00/5 · evidence 3%Figure 03 · Deployment 1.20/5 · evidence 1%1X NEO · Deployment 1.20/5 · evidence 3%Kepler Forerunner K3 · Deployment 1.00/5 · evidence 0%Kepler Forerunner K2 · Deployment 1.20/5 · evidence 0%NEURA 4NE-1 · Deployment 0.40/5 · evidence 3%Apptronik Apollo · Deployment 1.00/5 · evidence 0%Tesla Optimus · Deployment 0.60/5 · evidence 3%Boston Dynamics Atlas (electric) · Deployment 1.80/5 · evidence 2%Unitree H2 · Deployment 0.80/5 · evidence 4%Fourier GR-2 · Deployment 0.80/5 · evidence 3%XPeng IRON · Deployment 1.00/5 · evidence 3%Agibot A2 · Safety 1.60/5 · evidence 11%UBTech Walker S · Safety 0.80/5 · evidence 14%Figure 03 · Safety 1.00/5 · evidence 5%1X NEO · Safety 0.80/5 · evidence 15%Kepler Forerunner K3 · Safety 0.20/5 · evidence 5%Kepler Forerunner K2 · Safety 0.20/5 · evidence 5%NEURA 4NE-1 · Safety 0.60/5 · evidence 8%Apptronik Apollo · Safety 0.60/5 · evidence 5%Tesla Optimus · Safety 0.40/5 · evidence 33%Boston Dynamics Atlas (electric) · Safety 0.40/5 · evidence 33%Unitree H2 · Safety 0.20/5 · evidence 5%Fourier GR-2 · Safety 0.20/5 · evidence 15%XPeng IRON · Safety 0.20/5 · evidence 0%Rim = 2.0 of 5 · framed to the cohort's own ceiling0

3. Matrix view — index matrix

Detailed table of 13 offerings across 8 axes. Expand any axis to see 5 sub-axes.

Sort
MachineScore
1Agibot A2Agibot (Zhiyuan Robotics)1.50+0.1013% ev
1.20
27%
1.20
11%
0.80
25%
1.60
14%
2.00+0.80
8%
2.00
16%
1.60
4%
1.60
11%
2UBTech Walker SUBTech1.488% ev
1.60
12%
1.60
8%
1.60
6%
1.80
3%
0.40
13%
2.00
11%
2.00
3%
0.80
14%
3Figure 03Figure AI1.35+0.159% ev
0.80
14%
1.40+0.40
5%
1.60+0.20
16%
1.60+0.40
9%
1.60+0.20
12%
1.60
8%
1.20
1%
1.00
5%
41X NEO1X1.33+0.048% ev
1.60+0.40
15%
0.80
4%
1.60
7%
1.00
4%
2.00
8%
1.60
11%
1.20
3%
0.80
15%
5Kepler Forerunner K3Kepler Robotics1.23+0.067% ev
1.60
8%
1.20
8%
1.40
9%
1.20
5%
1.60+0.40
12%
1.60
5%
1.00
0%
0.20
5%
6Kepler Forerunner K2Kepler Robotics1.18+0.077% ev
1.60
8%
1.20
12%
1.40
8%
1.40+0.40
8%
0.80
6%
1.60
5%
1.20+0.20
0%
0.20
5%
7NEURA 4NE-1NEURA Robotics1.008% ev
0.80
10%
1.20
4%
1.00
8%
1.20
8%
1.20
11%
1.60
9%
0.40
3%
0.60
8%
8Apptronik ApolloApptronik0.98+0.018% ev
0.80
11%
1.20
13%
0.60
12%
1.60
8%
0.80
10%
1.20
6%
1.00
0%
0.60
5%
9Tesla OptimusTesla0.78+0.0523% ev
0.80
34%
0.80
8%
0.60
33%
1.00
23%
1.20+0.40
18%
0.80
36%
0.60
3%
0.40
33%
10Boston Dynamics Atlas (electric)Boston Dynamics0.7516% ev
0.80
26%
1.20
12%
0.40
33%
0.20
15%
0.80
19%
0.40
33%
1.80
2%
0.40
33%
11Unitree H2Unitree0.709% ev
1.20
9%
1.00
6%
0.80
11%
0.40
3%
0.20
5%
1.00
14%
0.80
4%
0.20
5%
12Fourier GR-2Fourier Intelligence0.57−0.0110% ev
0.80
14%
0.80
8%
1.00
13%
0.40
8%
0.20
5%
0.40
13%
0.80
3%
0.20
15%
13XPeng IRONXPeng0.57+0.095% ev
0.40
8%
0.60
2%
1.20
7%
0.40+0.40
8%
0.40
8%
0.40
3%
1.00+0.40
3%
0.20
0%

Each cell shows the evidence-backed read, its movement since last edition, and how much of the record supports it. The solid bar is the proven read; the pale bar behind is the analyst read — the gap between them is the proof gap. A faded bar means thin proof, not weak capability. Click any axis heading to open its five sub-axes.

4. Relative vs absolute view

How the same market looks on a relative lens against the true 0–5 scale.

Relative (market)zoomed to cohort
Agibot A2 · capability 1.50 · evidence 13%UBTech Walker S · capability 1.48 · evidence 8%Figure 03 · capability 1.35 · evidence 9%1X NEO · capability 1.33 · evidence 8%Kepler Forerunner K3 · capability 1.23 · evidence 7%Kepler Forerunner K2 · capability 1.18 · evidence 7%NEURA 4NE-1 · capability 1.00 · evidence 8%Apptronik Apollo · capability 0.98 · evidence 8%Tesla Optimus · capability 0.78 · evidence 23%Boston Dynamics Atlas (electric) · capability 0.75 · evidence 16%Unitree H2 · capability 0.70 · evidence 9%Fourier GR-2 · capability 0.57 · evidence 10%XPeng IRON · capability 0.57 · evidence 5%0.401.67262

Differences between machines are visible because the frame is the market itself.

Absolute (0–5)true 0–5 scale
Agibot A2 · capability 1.50 · evidence 13%UBTech Walker S · capability 1.48 · evidence 8%Figure 03 · capability 1.35 · evidence 9%1X NEO · capability 1.33 · evidence 8%Kepler Forerunner K3 · capability 1.23 · evidence 7%Kepler Forerunner K2 · capability 1.18 · evidence 7%NEURA 4NE-1 · capability 1.00 · evidence 8%Apptronik Apollo · capability 0.98 · evidence 8%Tesla Optimus · capability 0.78 · evidence 23%Boston Dynamics Atlas (electric) · capability 0.75 · evidence 16%Unitree H2 · capability 0.70 · evidence 9%Fourier GR-2 · capability 0.57 · evidence 10%XPeng IRON · capability 0.57 · evidence 5%0.005.001000

The same cohort against the full published scale — the field is still early.

Lens in use

Zoomed to the live cohort: capability 0.57–1.50 of 0–5, evidence 5–23% of 0–100%. Differences are magnified, not absolute.

Current market window

00.57–1.505