Head to head
Aria vs Sophia
Aria (Realbotix, from $125,000) against Sophia (Hanson Robotics, no public price), side by side on price, realism, specifications and availability.
The short version
- Aria scores higher on our Realism Index: 64/100 to Sophia's 40. The widest gap is body & skin (80 vs 32).
- Aria has a public price ($125,000); Sophia doesn't.
- Aria is no longer listed by its maker; Sophia is still a concept.
- Aria also carries a $199.99/month software subscription.
Composed from the two directory records, so it updates when they do.
Aria64/100
Sophia40/100
Realism Index sub-scores — six weighted dimensions built from public evidence. How we score
![]() | ![]() | |
|---|---|---|
| Status | Unlisted | Concept |
| Type | Companion | Humanoid |
| Form | Full body | Torso |
| Presents as | Female-presenting | Female-presenting |
| Price | Last listed Sept 2026: F-Series from $125k, M-Series from $95k, bust from $20k, plus $199.99/mo AI (launched at $175k). No longer on Realbotix's site, so confirm with Realbotix. The Aria persona is now also sold on Echo: $11,950 tabletop or $59,950 full body, deliveries H2 2027. | Not sold to consumers — a research and exhibition platform. The desktop Little Sophia retails around $149. |
| Subscription | $199.99/mo | — |
| Availability | Last listed Sept 2026 | Operational since 2016 |
| Origin | USA | Hong Kong |
| Realism index | 64/100 | 40/100 |
| Height | 175 cm · 5 ft 9 in (F-Series, third-party) | ~167 cm · 5 ft 6 in (with legs; usually shown as a torso on a base) |
| Weight | ~54 kg (F-Series, third-party) | — |
| Skin & materials | Platinum-cure silicone over an articulated frame | Frubber face and neck; transparent rear skull; decorative body |
| Actuation | Electric servos. F-Series: 17 facial motors, up to 44 DOF; Echo Full Body: 39 DOF | 30+ motors in the head and face |
| Power | Built-in battery | — |
| Runtime | 4–8 hours (F-Series) | — |
| AI & conversation | Proprietary deep-learning conversation; AI-agnostic (cloud GPT or on-device) | Hybrid — scripted dialogue, a chatbot, OpenCog reasoning and LLMs; neural-net face and emotion recognition |
| Voice | Customisable voices | Synthesised speech; primarily English |
| Mobility | Wheeled base — does not walk | Very limited — static or wheeled; a legged variant has been demoed but barely walks |
| Expressions | 17 facial motors; emotionally aware, tone-matching | 60+ facial expressions; gaze tracking |
| Languages | 100+ languages | Primarily English; multilingual demos |
| Connectivity | Wi-Fi, Bluetooth | Wi-Fi; cameras in the eyes and chest |


