Model reference · open weights
LongLive2.0-Robot-S is an open-weight video model from Efficient-Large-Model. LongLive2.0-Robot-S (BF16) weighs 20.0 GB; the smallest configuration that runs it is RTX 4090 24 GB.
For self-hosting, compare this build's requirements below with the GPU you plan to rent.
Summary of the Efficient-Large-Model/LongLive2.0-Robot-S model card, 2026-10-07
What it is
| Released by | Efficient-Large-Model |
|---|---|
| Released | 2026-10-06 |
| VRAM | 20.0 GB for the weights |
What it runs on
| Card | Weights | Memory |
|---|---|---|
| RTX 3060 12 GB … RTX 4060 Ti 16 GB | does not fit | |
| RTX 3090 24 GB | tight | 23.4 GB |
| RTX 4090 24 GB | tight | 23.4 GB |
| RTX 5090 32 GB | fits | 31.0 GB |
| L40S 48 GB | fits | 44.0 GB |
| A100 80 GB | fits | 78.2 GB |
| H100 80 GB | fits | 78.1 GB |
| RTX PRO 6000 Blackwell 96 GB | fits | 93.8 GB |
| DGX Spark (GB10) 128 GB unified | fits | 107 GB |
| H200 141 GB | fits | 138 GB |
| B200 180 GB | fits | 176 GB |
From the model card
Autoregressive robot-video generation weights. S is trained on short sequences of up to a few seconds. S and M describe training sequence length, not model size.
hf download Efficient-Large-Model/LongLive2.0-Robot-S --local-dir ./weights/LongLive2.0-Robot-S
Instantiate the matching LongLive 2.0 generator, then load the FP32 weights:
import torch
state_dict = torch.load("model.pt", map_location="cpu", mmap=True, weights_only=True)
model.load_state_dict(state_dict, strict=True)
Run the loading example from the downloaded directory. The matching inference runtime, VAE and text encoder are required separately; this is not a standalone pipeline. Optimizer and other training states are not included.
Original upstream model terms continue to apply.
Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.
How it works
Running it yourself
Rent a machine by the hour. How to run this model is on its model card.