TacitVLA: Reasoning Vision-Language-Action Models Without Reasoning Latency
TacitVLA: Reasoning Vision-Language-Action Models Without Reasoning Latency
TacitVLA is a vision-language-action model that improves performance without the latency of explicit reasoning. By decoupling action generation from reasoning, it achieves up to 3.4x lower latency while maintaining performance on tasks like manipulation and autonomous driving. It outperforms prior results on several benchmarks, including RoboMME and RMBench.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- code: not_reported
- weights: not_reported
- dataset: not_reported
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
sentation. We introduce TacitVLA, a reasoning VLA whose action expert never attends to the reasoning it generates for the current call. On long-horizon tasks where the reasoning text also serves as episodic memory, it can generate each entry during action execution rather than before acting. We evaluate it in a controlled study on four benchmarks spanning manipulation and autonomous driving, memory and non-memory tas
Open source S3
ks: RoboMME, RMBench, LIBERO, and PhysicalAI-AV. On all four, TacitVLA matches the performance of its explicit-reasoning counterpart at up to 3.4x lower latency, and matches the latency of the same architecture trained without reasoning at up to +17.1% success. In absolute terms, it surpasses the best previously published results on the RoboMME Counting suite (82.7% vs. 66.8%) and RMBench (77.5% vs. 56.5%), and reach
Open source S4
es 72.9% on LIBERO-Plus and a 52.5% scene score on the public split of the PhysicalAI-AV closed-loop driving challenge. Project page: https://tacitvla.github.io.
Open source S5
Source：arXiv Robotics — research abstracts · arxiv.org