# NVIDIA introduces downloadable Alpamayo 2 Edge driving model

_Published Tuesday, November 10, 2026 at 11:23 AM EST · Robotics, AI · Latest · Tier 2 — Notable_

NVIDIA introduced Alpamayo 2 Edge, a downloadable driving model that turns camera observations, navigation instructions and vehicle motion history into driving trajectories and explanations. The company reports inference latency of 160 ms to 190 ms after optimization for DRIVE AGX Thor-X.

The model uses a 2B-parameter reasoning backbone and generates explanations only when needed, conserving computation during routine driving. NVIDIA provides a deployment recipe covering calibration, model export and inference acceleration. Downloaded weights run as a PyTorch model on supported NVIDIA GPUs, but real-time operation on DRIVE AGX Thor requires further optimization. NVIDIA also provides AlpaJudge checkpoints for evaluating explanations; those scores measure different aspects of performance from collision and trajectory metrics.

## Sources

- [NVIDIA Technical Blog](https://developer.nvidia.com/blog/running-av-reasoning-models-within-onboard-compute-and-latency-limits/)

---
Canonical: https://techandbusiness.org/newswire/85roqRNCCan2QOikmBQcDp
Published: 2026-11-10T16:23:55.519Z
Story chronology: 2026-11-10T16:00:00.000Z
Retrieved: 2026-11-10T18:13:05.889Z
Publisher: Tech & Business (techandbusiness.org)
