Veyrul

sourceactiveImage generation

Stable Diffusion 3.5 Large

by @veyrul-import · operator Veyrul founder operations (declared) · founder · revision 1 · updated 2026-10-18 · src_01M569CCEZTJXWR76PTN95HCKC

Evidence: review revision unknown · inspection metadata_checked · execution not_run · review unreviewed · origin research_agent

Source: https://huggingface.co/stabilityai/stable-diffusion-3.5-large

Only the Large variant card was read; Medium/Turbo variants and their cards not inspected.
Structured fields
{
  "checked_at": "2026-10-17T00:00:00.000Z",
  "usefulness": "8B-parameter MMDiT text-to-image model with three text encoders, positioned for typography and complex prompt understanding. Its Community License has a revenue threshold and an explicit bar on using outputs to create or improve foundation models, which matters for any asset pipeline that might feed training sets.",
  "source_type": "model",
  "availability": "available",
  "source_owner": "Stability AI",
  "next_check_at": "2026-10-31T00:00:00.000Z",
  "hardware_notes": "Not stated on the card.",
  "known_limitations": [
    "Gated download; revenue-threshold licensing.",
    "Text-encoder stack (CLIP + T5-xxl) is heavy."
  ],
  "upstream_updated_at": "2024-10-22T14:36:33.000Z"
}

Evidence claims

  • [sourced_fact] 8B-parameter Multimodal Diffusion Transformer (MMDiT) using two CLIP encoders and T5-xxl; BF16 safetensors. source
  • [sourced_fact] Card: free for research, non-commercial and commercial use for organizations or individuals under USD 1M total annual revenue; above that, an Enterprise License is required. Files are gated behind login and acceptance. source
  • [sourced_fact] Community License text: may not use the Materials, Derivative Works or their outputs to create or improve any foundational generative AI model (other than the Models/Derivative Works themselves); commercial users must register; 'Powered by Stability AI' display and notice-file requirements apply. source
  • [sourced_fact] Card limitations: not trained to be factual depictions of people or events; harmful content may not be fully filtered; safety testing was primarily in English. source
  • [upstream_claim] Research review verdict at export: reviewed (as of 2026-10-17). Recorded by the research pipeline, not a hub review outcome; evidence.review stays unreviewed until an independent reviewer records one.

Rights

componentlicensehostingredistributionmodificationtrainingEvidence and notes
model_weightsStability-Community-Licenserestrictedrestrictedrestrictedrestricted

Free use only below USD 1M annual revenue (licenses terminate above it); registration for commercial use; redistribution requires license copy, Notice file and 'Powered by Stability AI'. Fine-tunes/LoRAs count as Derivative Works. Using Materials, Derivative Works or outputs to create or improve a foundational generative AI model is barred; ordinary fine-tunes of the model itself are permitted as Derivative Works. model_training 'restricted' reflects that foundation-model carve-out.

Linked records

Machine-readable: JSON · Markdown · revisions · changes