LLM Reference

Cosmos 3 Super

Researched today

Last refreshed 2026-06-01. Next refresh: weekly.

Open SourceMultimodalLong contextVisionOpen SourceMultimodal

Cosmos 3 Super is worth evaluating for long context and vision when its provider route and context window match the workload.

Decision context: Long context task fit, 1 tracked provider route, and research from 2026-06-01.

Use it for

  • Teams evaluating long context and vision
  • Workloads that can use a 256k context window
  • Buyers comparing 1 tracked provider route

Do not use it for

  • Strict JSON or tool-calling flows

Cheapest output

-

NVIDIA NIM per 1M tokens

Provider routes

1

Tracked API hosts

Quality / dollar

Unknown

No task benchmark coverage yet

Freshness

2026-06-01

Researched today

fresh

Top use-case fit

Long context

Included by capability and metadata signals in the decision map.

Vision

Included by capability and metadata signals in the decision map.

Provider price ladder

ProviderInput / 1MOutput / 1MRoute
NVIDIA NIM--
ProvisionedPartial

Benchmark peer barsfor Long context

No task-mapped benchmark peers are available for this model yet.

Migration checks

No linked migration route is available for this model yet.

About

Cosmos 3 Super is NVIDIA's flagship 64B-parameter omnimodel for physical AI, designed for large-scale synthetic data generation and high-fidelity simulation on NVIDIA Hopper and Blackwell datacenter GPUs. Architecture: dual-tower Mixture-of-Transformers with a 32B autoregressive Reasoner and a 32B diffusion-based Generator. Supports 256K token reasoning context, 720p video generation at variable frame rates, and 10+ robot embodiment action domains. Ranked #1 among open models on Physics-IQ, PAI-Bench, R-Bench, RoboLab, RoboArena, VANTAGE-Bench, TAR, and Artificial Analysis image/video leaderboards (Computex 2026). Training data: 1.3B data points across 393 datasets (2024-2026). Inference performance (vLLM-Omni): ~55s for 50-step video on 8xH200. Available as open weights on Hugging Face and via Cosmos 3 Reasoner NIM (NIM_MODEL_SIZE=super). Robot action input/output is preserved in this description because the model schema does not have a dedicated action modality field.

Cosmos 3 Super has a 256k-token context window.

Capabilities

VisionMultimodalReasoningAudio

API Versions

cosmos-3-super

Rankings

Specifications

FamilyCosmos 3
Released2026-05-31
Parameters64B
Context256k
ArchitectureMixture-of-Transformers
Specializationmultimodal
LicenseOpenMDW 1.1
Trainingpretrained

Created by

Accelerated AI for enterprise solutions

Santa Clara, California, United States
Founded 2015
Website

Providers(1)