dreamer_v3_400m(pretrained: bool = False, overrides: object = {})Construct the largest rung — no objectives, just the model.
A model dimension of 1536. The top of the scaling study, where the paper reports that larger models not only score higher but reach a given score in fewer environment steps — the data efficiency improves with capacity rather than trading against it.
Model Size
Parameters
pretrainedbool= FalseNo weights are published for this family; passing
True raises.**overridesobject= {}Optional
DreamerV3Config field overrides.Returns
DreamerV3ModelThe trunk at 400M parameters.
Notes
Reference: Hafner, Pasukonis, Ba, and Norouzi, "Mastering Diverse Domains through World Models", Nature 640 (2025), 647-653 (arXiv:2301.04104), Figure 6c and Table 3.
Examples
>>> from lucid.models import dreamer_v3_400m
>>> model = dreamer_v3_400m(action_dim=6).eval()
>>> model.config.deter_size, model.config.hidden_size
(12288, 1536)