Hardware
Generate Trajectories, Reasoning Traces, and Auto-Labels with NVIDIA Alpamayo 2 Super
NVIDIA Alpamayo 2 Super is a publicly available 34‑billion‑parameter vision–language–action model that merges a 32‑B Cosmos 3 Super Reasoner with a 2‑B action‑expert diffusion network. It produces uni
NVIDIA Alpamayo 2 Super is a publicly available 34‑billion‑parameter vision–language–action model that merges a 32‑B Cosmos 3 Super Reasoner with a 2‑B action‑expert diffusion network. It produces unified vehicle trajectories, Chain‑of‑Causation reasoning traces, high‑level meta‑actions, 360° perception over up to seven cameras, and automatically labeled scene grounds for autonomous‑vehicle development. In benchmark tests it outperforms prior models with minimum ADE 6 of 0.91 m, AV reasoning accuracy 0.43, LingoQA 79.2%, meta‑action IoU exceeding 70% in lateral/longitudinal/lane scenarios, and a closed‑loop AlpaSim score of 1.50, enabling scalable, end‑to‑end AV training and evaluation.
Related
- How to Post-Train Autonomous Vehicle Models in Closed-Loop with NVIDIA Alpamayo
- NVIDIA Alpamayo 2 Super, the Frontier Open Model for Robotaxis and Autonomous Vehicles, Now Available for Commercial Use
- Develop Physical AI Reasoning, World, and Action Models with NVIDIA Cosmos 3
Source: NVIDIA Developer | 2026-08-04