NVIDIA Alpamayo 2 Super
Last reviewed
Jun 2, 2026
Sources
7 citations
Review status
Source-backed
Revision
v1 · 1,788 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Jun 2, 2026
Sources
7 citations
Review status
Source-backed
Revision
v1 · 1,788 words
Add missing citations, update stale details, or suggest a clearer explanation.
NVIDIA Alpamayo 2 Super is an open, 32-billion-parameter reasoning-based vision-language-action model (VLA) for safe, Level 4 robotaxi and autonomous-vehicle development, announced by NVIDIA at GTC Taipei during COMPUTEX 2026 on June 1, 2026. It is the flagship member of the open Alpamayo family, which NVIDIA first introduced at CES in January 2026. Alpamayo 2 Super roughly triples the parameter count of the original 10-billion-parameter Alpamayo 1 model and is designed to perceive a vehicle's full 360-degree surroundings, reason step by step about complex driving situations, and explain the decisions it makes. NVIDIA positions it as a "teacher" model that developers can fine-tune and distill into smaller networks that run in real time on in-vehicle hardware.[1][2]
Alpamayo, named after a peak in the Peruvian Andes, is built on the NVIDIA Cosmos family of world foundation models. Alongside Alpamayo 2 Super, NVIDIA announced a set of companion tools meant to complete the pipeline from real-world data capture to closed-loop training and on-road deployment, including the AlpaGym reinforcement learning framework, the OmniDreams generative world model, and new NVIDIA Omniverse NuRec neural-reconstruction models. Code and model weights are expected to be released in summer 2026 on GitHub and Hugging Face.[1][2]
| Attribute | Detail |
|---|---|
| Full name | NVIDIA Alpamayo 2 Super |
| Developer | NVIDIA |
| Announced | June 1, 2026, at GTC Taipei (COMPUTEX 2026) |
| Model type | Reasoning-based vision-language-action (VLA) model |
| Parameters | 32 billion |
| Foundation | Built on NVIDIA Cosmos world foundation models |
| Target use | Level 4 robotaxi and autonomous-vehicle development |
| Predecessor | Alpamayo 1 (10 billion parameters, CES 2026) |
| Companion tools | AlpaGym, OmniDreams, Omniverse NuRec, AlpaSim |
| Release plan | Inference code on GitHub, weights on Hugging Face, summer 2026 |
| Deployment target | Distilled onto NVIDIA DRIVE AGX Thor within DRIVE Hyperion |
| License | Open model weights with open-source inference and post-training scripts |
NVIDIA launched the Alpamayo family at CES on January 5, 2026, describing it as the first open, large-scale collection of reasoning VLA models, simulation frameworks, reinforcement learning infrastructure, and physical-AI datasets for autonomous driving. The goal was to give the global robotaxi and AV industry a shared, transparent foundation so that developers would not have to build core autonomy infrastructure from scratch.[2][3]
The original Alpamayo 1 is a 10-billion-parameter chain-of-thought, reasoning-based VLA model. Rather than mapping camera pixels directly to steering and braking commands the way a perception-only stack does, Alpamayo takes in multi-camera video, navigation inputs, and driving context, then generates a driving trajectory together with a reasoning trace that explains the choice. NVIDIA calls these traces Chain-of-Causation (CoC) reasoning. The approach lets a vehicle think through a novel or rare scenario step by step, for example deciding how to proceed through a busy intersection during a traffic-light outage it has never encountered before.[2][3]
The technical foundation for the family was described in the research paper "Alpamayo-R1: Bridging Reasoning and Action Prediction for Generalizable Autonomous Driving in the Long Tail," posted to arXiv by an NVIDIA team in late 2025. That work pairs Cosmos-Reason, a vision-language model pre-trained for physical AI, with a diffusion-based trajectory decoder that produces dynamically feasible trajectories in real time. It also introduces the Chain-of-Causation dataset, built through a hybrid auto-labeling and human-in-the-loop pipeline that produces decision-grounded, causally linked reasoning traces aligned with driving behaviors. The paper reported that scaling the model from 0.5 billion to 7 billion parameters produced consistent gains, foreshadowing the much larger Alpamayo 2 Super.[4]
By the time Alpamayo 2 Super was announced, NVIDIA said the Alpamayo models had been downloaded close to 400,000 times. The Alpamayo lineup also won a COMPUTEX 2026 Best Choice Award in the vehicle technology category.[1][2]
The "Super" suffix marks Alpamayo 2 as the high-capacity, top-of-range member of the second generation, in contrast to the smaller Nano variants NVIDIA offers for on-device use. The most concrete difference from Alpamayo 1 is scale: 32 billion parameters versus 10 billion, roughly a threefold increase. NVIDIA frames the larger model as a teacher that captures the broadest possible driving knowledge and reasoning ability, which developers then distill into compact student models small enough to run inside a vehicle.[1][2]
Beyond raw size, Alpamayo 2 Super adds several capabilities aimed squarely at the hardest cases in autonomous driving:
| Capability | What it adds |
|---|---|
| Full-surround perception | Expands from front-focused cameras to 360-degree situational awareness across front, side, and rear views |
| Meta-Actions | Adds macro driving actions such as yield, lane change, and stop to the model's output |
| Reasoning auto-labeling with 2D grounding | Generates high-quality reasoning labels tied to image regions, compressing annotation cycles from months to days |
| Improved Chain-of-Causation traces | Produces better reasoning and trajectories in rare, complex, long-tail scenarios |
| 3D spatial understanding | Strengthens spatial reasoning and trajectory prediction in difficult conditions |
The full-surround upgrade is significant because earlier reasoning was concentrated on what the vehicle could see ahead. Extending it to a 360-degree view lets the model reason about merging traffic, vehicles approaching from the side, and events behind the car. The Meta-Actions output gives the model a vocabulary of high-level maneuvers, which makes its intentions easier to interpret and validate. NVIDIA emphasizes that this interpretability is meant to support safety validation and collaboration with regulators, since the model can verbalize why it chose a given action.[1][2]
Alpamayo 2 Super was announced alongside a toolchain that NVIDIA describes as completing the path from real-world data capture to closed-loop training and in-vehicle deployment.[1][2]
AlpaSim is NVIDIA's open-source, end-to-end closed-loop simulation framework for autonomous vehicles, first released with the Alpamayo family at CES. It validates model decisions against simulated real-world consequences and provides realistic sensor modeling, configurable traffic dynamics, and scalable closed-loop testing.[2][3]
AlpaGym is an open-source, high-throughput, closed-loop reinforcement learning framework. It runs models through continuous cycles of decision and observation inside the AlpaSim environment, which exposes compounding errors and edge-case failures that static, pre-recorded datasets tend to miss. Training a driving policy in this loop, rather than only on fixed clips of past driving, is meant to make the model more robust when small mistakes would otherwise accumulate.[1][2]
OmniDreams is a generative world model for photorealistic, closed-loop AV scenario generation. It lets developers simulate rare and long-tail driving scenarios at scale, generating the kinds of unusual situations that are hard to capture often enough in real-world fleet data.[1][2]
The new NVIDIA Omniverse NuRec models add a neural-reconstruction skill that turns real-world fleet recordings into photorealistic 3D scenes. Those reconstructed scenes can then be replayed and modified in simulation, and adapted across different vehicle sensor configurations, so that data collected by one fleet can be reused to train and test other vehicles.[1][2]
Taken together, the tools give what NVIDIA calls a continuous path from open-loop pretraining on recorded data to closed-loop refinement in simulation. The Chain-of-Causation auto-labeling pipeline is also released as open source, which lets developers generate their own causally grounded reasoning labels rather than annotating them by hand.[1][2]
Alpamayo 2 Super sits at the software layer of NVIDIA's broader autonomous-vehicle platform. Because a 32-billion-parameter model is far too large to run directly on a vehicle, NVIDIA's intended workflow is to use Alpamayo 2 Super as a teacher and distill it into compact models that run on NVIDIA DRIVE AGX Thor, the in-car compute platform, within the DRIVE Hyperion reference architecture for sensors and compute. The simulation and reconstruction tooling is built on the AlpaSim microservice simulation stack together with Omniverse NuRec.[1][2]
The model's foundation in NVIDIA Cosmos connects it to NVIDIA's wider physical-AI strategy, since Cosmos world foundation models underpin both the reasoning core (via Cosmos-Reason) and the generative simulation used to train and test it. In this way Alpamayo spans NVIDIA's data, simulation, and on-vehicle layers as a single open ecosystem rather than a standalone model.[1][2][4]
NVIDIA announced Alpamayo 2 Super as an open release. Following the precedent set by Alpamayo 1, which shipped with open model weights and open-source inferencing scripts on Hugging Face, Alpamayo 2 Super is expected to be available in summer 2026, with inference code on GitHub and model weights on Hugging Face, accompanied by post-training scripts. The companion AlpaGym framework and the Chain-of-Causation auto-labeling pipeline were also described as open source.[1][2][3]
As of the GTC Taipei announcement, NVIDIA had not published a specific calendar date, a formal license name, or quantitative benchmark results for Alpamayo 2 Super. The public claims to date are architectural, describing improved reasoning, 3D spatial understanding, and trajectory prediction in long-tail scenarios rather than scored results on a named benchmark.[1][2]
Coverage of the launch focused on the model's size and on the open, full-stack nature of the Alpamayo ecosystem. Trade and financial press, including GamesBeat and reporting carried through GlobeNewswire and Yahoo Finance, described Alpamayo 2 Super as a 32-billion-parameter step up from the 10-billion-parameter first generation and highlighted the AlpaGym and OmniDreams tools and the planned summer release on GitHub and Hugging Face. Several outlets framed it as part of NVIDIA's push to give the entire robotaxi industry a common, interpretable foundation for Level 4 autonomy.[5][6][7]
In NVIDIA's keynote, founder and CEO Jensen Huang summarized the pitch: "Alpamayo is the moment cars begin to safely reason, not just drive. Only NVIDIA makes available open models, simulation, real-world data and agent skills so the entire global robotaxi ecosystem can develop level 4 capabilities that understand edge cases, explain decisions, earn trust and scale safely to millions of vehicles."[1]