Gemma's team has unveiled "Reachy on Wheels," a mobile version of the Reachy Mini robot that uses on-device Gemma models to converse, listen and move through its surroundings without a cloud connection.
Gemma team · Reachy Mini on Wheels
A mobile robot that talks, listens and moves using only on-device Gemma models
The battery-powered Reachy Mini on Wheels handles speech, listening and autonomous navigation locally, with no cloud connection involved.
How the on-device system is organized
Speech
Conversation handled by an on-device Gemma model
→
Listening
Audio input processed on the robot itself
→
Navigation
Autonomous movement through its surroundings
Power and connectivity: battery-powered, running locally, with no cloud dependency.
Reported advantages of local-only
Data stays on the device, supporting privacy for personal audio and sensor data
Can reduce latency compared with round trips to remote servers
Limits the need for network connectivity
Trade-offs and open questions
On-device models generally have limits on size and reasoning depth versus large cloud systems
Battery life is not quantified in the material available
Navigation reliability in unfamiliar spaces is not quantified
No independent evaluations or cloud comparisons reported yet
Why it matters
The significance lies in combining speech, audio input and movement on one local device.
Usually split
These functions are typically spread across services, often with remote servers.
Small model, bigger role
A compact language model takes on tasks that would otherwise depend on cloud infrastructure.
Next steps
Further technical detail is needed to judge how broadly such designs can be applied.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article...