{"id":"alpamayo-2-edge","slug":"alpamayo-2-edge","name":"Alpamayo 2 Edge","orgId":"nvidia","family":"alpamayo-2-edge","releaseDate":null,"availability":"ga","openWeights":true,"license":"Model weights OpenMDW-1.1; source code Apache License 2.0 (model card and NVlabs/alpamayo2-edge)","paramsTotal":null,"paramsSource":null,"contextWindow":null,"knowledgeCutoff":null,"modelCardUrl":"https://developer.nvidia.com/blog/running-av-reasoning-models-within-onboard-compute-and-latency-limits/","hfRepo":null,"modalities":["robotics"],"aliases":["Alpamayo 2 Edge","alpamayo-2-edge","Alpamayo2-Edge","alpamayo 2 edge","nvidia alpamayo 2 edge","nvidia/Alpamayo2-Edge","NVlabs/alpamayo2-edge"],"aliasNote":null,"origin":"tracked","tracked":{"trackedSince":"2026-11-10","announced":"2026-11-10","blurb":"NVIDIA's 2B-parameter reasoning vision-language-action model in the Alpamayo family, built on the Nemotron-based backbone of NVIDIA Cosmos 3 Edge and paired with a diffusion action expert. It generates Chain-of-Causation reasoning only when needed and produces driving trajectories from surround-camera observations, temporal context, and navigation inputs. Released November 10, 2026 as part of the Alpamayo Open Platform, with public weights on Hugging Face under OpenMDW-1.1 and code under Apache License 2.0. Sized for real-time in-vehicle use on NVIDIA DRIVE AGX Thor.","api":"none","openWeights":"yes","params":{"text":"NVIDIA describes it as a 2B-parameter model. Hugging Face card: backbone 2.45B (1.96B language model, 0.41B SigLIP2 vision encoder, 0.08B projector) plus a 0.2B action expert","source":"NVIDIA blog and nvidia/Alpamayo2-Edge Hugging Face model card"},"facts":[{"label":"Type","value":"Reasoning vision-language-action (VLA) model with adaptive Chain-of-Causation reasoning and a diffusion action expert (NVIDIA blog and model card)"},{"label":"Base","value":"Built on the 2B Nemotron-based reasoning backbone of NVIDIA Cosmos 3 Edge (NVIDIA blog and model card)"},{"label":"Target hardware","value":"NVIDIA DRIVE AGX Thor (NVIDIA blog)"},{"label":"Inputs","value":"Multi-camera video, ego-motion history, and navigation instructions; optional free-form visual question answering (NVIDIA blog and model card)"},{"label":"Trajectory","value":"64 waypoints from 0.1 through 6.4 seconds at 0.1-second intervals (model card)"},{"label":"Training data","value":"Roughly 115,000 hours of multi-camera driving video and about 3,700,000 Chain-of-Causation traces (model card)"},{"label":"Weights","value":"Public, ungated Hugging Face repo nvidia/Alpamayo2-Edge (BF16 safetensors). Official card release date November 10, 2026"},{"label":"License","value":"Model weights OpenMDW-1.1; source code Apache License 2.0 (model card and NVlabs/alpamayo2-edge)"},{"label":"Code","value":"https://github.com/NVlabs/alpamayo2-edge"},{"label":"API","value":"No public cloud API announced"},{"label":"Related","value":"Separate from Alpamayo 2 Super (used as a teacher for CoC autolabels), AlpaJudge, Cosmos 3 Edge, and other Cosmos pages"}],"claims":["Lab claim: post-trained with closed-loop reinforcement learning in NVIDIA AlpaGym, improving closed-loop driving performance by 48.9% on top of the best-performing checkpoint without closed-loop RL (NVIDIA blog).","Lab claim: optimized pixel-to-trajectory inference latency of 160 ms to 190 ms on NVIDIA DRIVE AGX Thor, depending on the amount of reasoning (NVIDIA blog).","Lab claim: Lingo-Judge score of 75.2 on LingoQA (NVIDIA Hugging Face model card).","Lab claim: in AlpaSim closed-loop evaluation on 875 scenarios from PhysicalAI-AV-NuRec, with three rollouts per scenario, 15.15 ± 0.04 km driven per at-fault incident (NVIDIA Hugging Face model card).","Lab claim: open-loop minADE_6 at 5s of 0.52 m on 1,392 samples from the PhysicalAI-AV dataset (NVIDIA Hugging Face model card).","Lab claim: achieves approximately 40× the distance between at-fault incidents in AlpaSim compared with Qwen-Drive-1.0 (NVIDIA blog)."],"sources":[{"kind":"lab","title":"NVIDIA blog: Running AV Reasoning Models Within Onboard Compute and Latency Limits","url":"https://developer.nvidia.com/blog/running-av-reasoning-models-within-onboard-compute-and-latency-limits/","date":"2026-11-10"},{"kind":"repo","title":"nvidia/Alpamayo2-Edge on Hugging Face","url":"https://huggingface.co/nvidia/Alpamayo2-Edge","date":"2026-11-10"},{"kind":"repo","title":"NVlabs/alpamayo2-edge on GitHub","url":"https://github.com/NVlabs/alpamayo2-edge"}],"firstBenchmark":null},"organization":{"id":"nvidia","name":"NVIDIA","slug":"nvidia","country":null,"url":null,"color":"#76B900","mark":"N"},"price":null,"answer":{"word":"Pending benchmarks.","scope":"Announced Nov 10, 2026; no independent results yet.","lead":"As of Nov 22, 2026, Alpamayo 2 Edge is pending benchmarks on isitsota: announced Nov 10, 2026, with no independent results on the boards we track yet.","confidence":"low"},"status":"PENDING_BENCHMARKS","verdicts":[]}