Cosmos 3 Nano — VRAM, speed & local setup
Written by Jakub Rusinowski · Last updated
The mid-tier workhorse — a 16B-parameter two-tower MoT on a dense 8B backbone, tuned for high-quality video and action reasoning in a fraction of a second. Nano is the balance point: much cheaper to serve than Super while still leading open models (NVIDIA reports it matches the second-best result on R-Bench). Weights are ~32 GB at BF16, but NVIDIA recommends an RTX PRO 6000 / 96 GB-class GPU for full-context BF16 inference because video generation activations are memory-hungry. Omnimodal I/O (text, image, video, audio, action). OpenMDW-1.1. Specs from launch coverage and the Hugging Face model card — verify before relying on them.
Cosmos 3 Nano needs about 10 GB of VRAM at BF16 — quantized weights plus framework overhead, before any KV cache. On Apple Silicon that figure comes out of unified memory.
VRAM and speed by quantization
Quoted against NVIDIA RTX 4090 (24 GB). Weights plus overhead only: this model's architecture is unpublished, so no KV cache is added.
| Quant | Memory | VRAM | Speed (est.) | Fit |
|---|---|---|---|---|
| Q2_K 2.63 bpw | 6.1 GB | ~150 tok/s | Fits | |
| Q3_K_M 3.41 bpw | 7.6 GB | ~130 tok/s | Fits | |
| Q4_K_M 4.83 bpw | 10.5 GB | ~104 tok/s | Fits | |
| Q5_K_M 5.67 bpw | 12.1 GB | ~93 tok/s | Fits | |
| Q6_K 6.56 bpw | 13.9 GB | ~84 tok/s | Fits | |
| Q8_0 8.50 bpw | 17.8 GB | ~69 tok/s | Fits | |
| F16 16.00 bpw | 32.8 GB | ~6 tok/s | Offload |
Black marker = usable memory on the NVIDIA RTX 4090 (24 GB). Estimates from the memory-bandwidth roofline on the methodology page. Cosmos 3 Nano VRAM calculator →
Get Cosmos 3 Nano running
The cheapest catalogued GPU that runs Cosmos 3 Nano is the Intel Arc B580 (12 GB).
or compare on Vast.ai from $0.35/hr (typical low · varies)
As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.
How to run Cosmos 3 Nano
Install Ollama, then run:
ollama run cosmos-3Specifications
Preview — The model is released, but these specs are thin or rest on a single source. Individual fields may be wrong.
Preview — The model is released, but these specs are thin or rest on a single source. Individual fields may be wrong.
- Parameters
- 16B total (8B dense transformer backbone)
- Context window
- Omnimodal I/O (video/action)
- Architecture
- Two-tower Mixture-of-Transformers — autoregressive reasoner (VLM) + diffusion generator, shared multimodal attention
- Provider
- NVIDIA
- Licence
- OpenMDW-1.1
- Practical deployment
- Laptop
- Specified at
- BF16
- System RAM
- 96 GB
- Record updated
- 2026-07-21
Commercial use permitted. No usage restrictions beyond attribution.
Quality and use cases
Scores as published by the model’s authors or an independent evaluator — quality, not throughput, and not measured by us.
| Benchmark | Score | Provenance |
|---|---|---|
| R-Bench (Image-to-Video, open-source) | 2 / 2 2nd best (rank) | vendor-claimed · NVIDIA Cosmos 3 (launch) |
Can I run Cosmos 3 Nano on my GPU?
- Cosmos 3 Nano on AMD Radeon RX 7800 XT
- Cosmos 3 Nano on AMD Radeon RX 7900 XT
- Cosmos 3 Nano on AMD Radeon RX 9060 XT 16GB
- Cosmos 3 Nano on AMD Radeon RX 9060 XT 8GB
- Cosmos 3 Nano on AMD Radeon RX 9070
- Cosmos 3 Nano on AMD Radeon RX 9070 XT
- Cosmos 3 Nano on Apple M1
- Cosmos 3 Nano on Apple M1 Max
- Cosmos 3 Nano on Apple M2 Ultra
- Cosmos 3 Nano on Apple M3 Max
- Cosmos 3 Nano on Apple M5 Pro
- Cosmos 3 Nano on Intel Arc B570
- Cosmos 3 Nano on Intel Arc B580
- Cosmos 3 Nano on NVIDIA A100 80GB (PCIe)
- Cosmos 3 Nano on NVIDIA GeForce RTX 3060 (12GB)
- Cosmos 3 Nano on NVIDIA GeForce RTX 3070
- Cosmos 3 Nano on NVIDIA GeForce RTX 3070 Ti
- Cosmos 3 Nano on NVIDIA GeForce RTX 3080 (10GB)
- Cosmos 3 Nano on NVIDIA GeForce RTX 4060
- Cosmos 3 Nano on NVIDIA GeForce RTX 4060 Ti 16GB
- Cosmos 3 Nano on NVIDIA GeForce RTX 4070
- Cosmos 3 Nano on NVIDIA GeForce RTX 4070 Super
- Cosmos 3 Nano on NVIDIA GeForce RTX 4070 Ti
- Cosmos 3 Nano on NVIDIA GeForce RTX 4070 Ti Super
- Cosmos 3 Nano on NVIDIA GeForce RTX 4080
- Cosmos 3 Nano on NVIDIA GeForce RTX 4080 Super
- Cosmos 3 Nano on NVIDIA GeForce RTX 5060
- Cosmos 3 Nano on NVIDIA GeForce RTX 5060 Ti 16GB
- Cosmos 3 Nano on NVIDIA GeForce RTX 5060 Ti 8GB
- Cosmos 3 Nano on NVIDIA GeForce RTX 5070
- Cosmos 3 Nano on NVIDIA GeForce RTX 5070 Ti
- Cosmos 3 Nano on NVIDIA GeForce RTX 5080
- Cosmos 3 Nano on NVIDIA H100 80GB (PCIe)
- Cosmos 3 Nano on NVIDIA L40S
- Cosmos 3 Nano on NVIDIA RTX 6000 Ada Generation
Other Cosmos 3 sizes
Cosmos 3 Nano — frequently asked questions
How much VRAM does Cosmos 3 Nano need?
About 10 GB at BF16 — quantized weights plus framework overhead, before any KV cache. The cache grows with context length and is added on top; the table above folds it in. Apple Silicon counts unified memory toward the same figure.
Does Cosmos 3 Nano run on an RTX 4090 (24 GB)?
Yes. Cosmos 3 Nano needs about 10 GB at BF16, inside a 24 GB card, at an estimated 104 tokens/sec.
How do I run Cosmos 3 Nano locally?
Install Ollama and run `ollama run cosmos-3`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.
What other sizes does Cosmos 3 come in?
Cosmos 3 Super (39 GB), Cosmos 3 Nano (10 GB), Cosmos 3 Edge (3 GB). Every size shares the family's training and licence; the larger ones score higher and need proportionally more memory.