NVIDIA Launches Alpamayo 2 Edge for Onboard AV Reasoning

It uses closed-loop reinforcement learning to improve driving performance by 48.9% and integrates AlpaJudge for human-calibrated reasoning evaluation. The model supports adaptive reasoning, multi-resolution inputs, and speculative decoding with DFlash2, achieving 160–190 ms pixel-to-trajectory latency.

Summarised from the original article by an open-source model, then checked automatically: every number and English name must appear in the source. The original article is authoritative.