humanoidversus
Compare

Humanoid robot task evidence

Which humanoid robots have evidence for conversation?

Responding or interacting through speech, with any disclosed human control identified.

8 distinct robots · 8 records · library data last updated Oct 13, 2026

1X EVE →

Conversation

Spoken commands selecting and chaining short robot skills

1X EVE →
Date
May 31, 2024
Setting
Unknown
Evidence type
Maker demo
Evidence stated by
Maker
Stated autonomy
Mixed
Autonomy stated by
Maker
Playback speed
Unknown
Continuous take
Unknown

Outcome / scope: 1X states that humans decide when EVE switches tasks while learned policies execute the actions. This combines autonomous skills with human task sequencing rather than an autonomous long-horizon plan.

Metrics: no measurements recorded.

Evidence last verified Oct 13, 2026

1X NEO →

Conversation

Spoken contextual assistance

1X NEO →
Date
Oct 28, 2025
Setting
Home
Evidence type
Maker claim
Evidence stated by
Maker
Stated autonomy
Unknown
Autonomy stated by
No source states it
Playback speed
Unknown
Continuous take
Unknown

Outcome / scope: 1X describes built-in language-model conversation, visual context and memory. The announcement provides no independent conversation test.

Metrics: no measurements recorded.

Evidence last verified Oct 13, 2026

AgiBot A2 →

Conversation

Real-time full-duplex spoken dialogue

AgiBot A2 →
Date
Unknown
Setting
Unknown
Evidence type
Product spec
Evidence stated by
Maker
Stated autonomy
Unknown
Autonomy stated by
No source states it
Playback speed
Unknown
Continuous take
Unknown

Outcome / scope: The original A2 page describes language-model-driven dialogue and sound-source localization. It supplies no conversation trial protocol.

Metrics: no measurements recorded.

Evidence last verified Oct 13, 2026

AgiBot Genie G2 →

Conversation

Multi-user conversation with gaze tracking

AgiBot Genie G2 →
Date
Unknown
Setting
Unknown
Evidence type
Product spec
Evidence stated by
Maker
Stated autonomy
Unknown
Autonomy stated by
No source states it
Playback speed
Unknown
Continuous take
Unknown

Outcome / scope: AgiBot lists multi-user conversation and gaze tracking for G2. This feature listing does not document a measured conversation test.

Metrics: no measurements recorded.

Evidence last verified Oct 13, 2026

AgiBot Lingxi X2 →

Conversation

Asking shoppers about cookware needs and recommending products

AgiBot Lingxi X2 →
Date
Sep 28, 2026
Setting
Retail
Evidence type
Customer deployment
Evidence stated by
Maker
Stated autonomy
Unknown
Autonomy stated by
No source states it
Playback speed
Unknown
Continuous take
Unknown

Outcome / scope: AgiBot describes X2 asking a shopper about household size, cooking habits and budget. Store staff handle promotions and purchases; the source does not disclose task autonomy.

Metrics: no measurements recorded.

Evidence last verified Oct 13, 2026

Agility Digit →

Conversation

Accepting spoken natural-language task requests

Agility Digit →
Date
Dec 13, 2023
Setting
Unknown
Evidence type
Maker demo
Evidence stated by
Maker
Stated autonomy
Unknown
Autonomy stated by
No source states it
Playback speed
Unknown
Continuous take
Unknown

Outcome / scope: Agility describes an interactive LLM demo where people request tasks by speaking. This shows command-based interaction, rather than an independently tested conversational service.

Metrics: no measurements recorded.

Evidence last verified Oct 13, 2026

Fourier GR-3 →

Conversation

Spoken interaction with CES visitors

Fourier GR-3 →
Date
Jan 2026
Setting
Event
Evidence type
Maker claim
Evidence stated by
Maker
Stated autonomy
Unknown
Autonomy stated by
No source states it
Playback speed
Unknown
Continuous take
Unknown

Outcome / scope: The maker describes visitor conversation and touch-based interaction at its booth. It provides no task-specific autonomy disclosure or conversation evaluation.

Metrics: no measurements recorded.

Evidence last verified Oct 13, 2026

XPeng IRON →

Conversation

Conversation using XPENG's robot model stack

XPeng IRON →
Date
Nov 5, 2025
Setting
Unknown
Evidence type
Maker claim
Evidence stated by
Maker
Stated autonomy
Unknown
Autonomy stated by
No source states it
Playback speed
Unknown
Continuous take
Unknown

Outcome / scope: XPENG lists conversation among the next-generation IRON's model-enabled capabilities. The release supplies no conversation transcript or measured evaluation.

Metrics: no measurements recorded.

Evidence last verified Oct 13, 2026

Explore other task evidence

Full matrix and demo library →Walking →Running →Climbing stairs →Handling totes →Pick and place →Bin picking →Assembly →Machine tending →Kitting →Folding laundry →Loading a dishwasher →Cooking and food preparation →Tidying →Cleaning surfaces →Opening doors →Using tools →In-hand manipulation →Guiding and reception →Dance and performance →Sports →Inspection →Teleoperated avatar →Other tasks →