Robot AI model
TinyVLA
TinyVLA is a family of compact VLA policies that combines multimodal models with diffusion-based action generation. The authors evaluate single-arm Franka and bimanual UR5 manipulation and publish training code with several LLaVA-Pythia backbone sizes.
Records last verified Oct 13, 2026 · Send a correction
- Type
- VLA
- Released / introduced
- Sep 19, 2024
- Public weights
- Not established
- Licence / scope
- Not established
- Parameters
- Not disclosed
- Base model
- LLaVA-Pythia (size depends on variant)
- Humanoid association
- None reported
- Inputs
- vision, language
- Outputs
- continuous robot actions
Release dates can mark an announcement or paper introduction. Public checkpoints, code licences and weight terms are distinct; follow the publisher’s links for the stated scope.
Documented embodiments
Source-associated hardware; this list does not establish plug-and-play compatibility or commercial robot capability.
- Franka robot arm
Single-arm evaluation.
- Bimanual UR5
Bimanual evaluation.
More robot AI models
Evidence
Sources
- [s1] TinyVLA: Towards Fast, Data-Efficient Vision-Language-Action Models for Robotic Manipulation ↗
TinyVLA research team · paper · published Sep 19, 2024 · accessed Oct 13, 2026
- [s2] TinyVLA project: affiliations and robot evaluations ↗
TinyVLA research team · official · accessed Oct 13, 2026
- [s3] TinyVLA official README: backbone variants and code release ↗
TinyVLA research team · official · accessed Oct 13, 2026