Skip to content
OpenMayhem
ExplorePlaygroundFAQ
GitHubGet started
OpenMayhem
ExplorePlaygroundFAQ
GitHubGet started
  1. Explore
  2. /
  3. Sulphur-2
S

SulphurAI

Video model

Sulphur-2

An uncensored LTX-2.3 video model that generates motion and synchronized audio.

Video
Use this model See capabilities

Input to output

Text + optional image→Video + audio
Modes
T2V + I2V
Model size
LTX-2.3 base
Capabilities
Video
License
Not declared

Useful features

Text to videoImage to videoGenerated audio

Capabilities

What you can use it for.

The most useful reasons to choose this model, without making you read through its repository first.

Text to videoImage to videoGenerated audioFewer refusals
01

Text-to-video

02

Image animation

03

Video with generated audio

For providers

Want to serve this model?

The provider requirements live here, separate from the information you need to choose and use the model.

Starting point48 GB GPU memory
Check your hardware in the fit calculator

48 GB CUDA + 64 GB RAM

FP8 / quantized + offload · Diffusers / ComfyUI

Needs validation

A quantized and offloaded route may fit, but OpenMayhem should benchmark resolution, frame count, audio length, and concurrency before offering it.

80 GB CUDA memory

BF16 / FP8 · Diffusers / ComfyUI

Conservative estimate

Video frames, the LTX-2.3 model stack, VAE, and audio generation need substantial working memory; 80 GB is the credible provider target.

Source and license

Check the original model.

OpenMayhem links back to the source repository so you can inspect the model card, files, license, limitations, and creator guidance directly.

Source repositorySulphurAI/Sulphur-2-baseOpen model card
OpenMayhem

An open, peer-to-peer marketplace for AI inference. Buy compute through your own gateway, or sell the compute you already own.

Project

  • Explore models
  • Hardware checker
  • Trust & fees
  • GitHub repository ↗ (opens in a new tab)
  • Marketplace rules ↗ (opens in a new tab)
  • Model catalog ↗ (opens in a new tab)

Ecosystem

  • Trac Network ↗ (opens in a new tab)
  • Trac Systems ↗ (opens in a new tab)
  • Community on Reddit ↗ (opens in a new tab)
  • Community on Telegram ↗ (opens in a new tab)

OpenMayhem is built and operated by Trac Systems UG (haftungsbeschränkt). The software is open source and released under the MIT license; the OpenMayhem and Trac names and logos are trademarks of Trac Systems. The repository is the source of truth for everything described here.

Legal noticePrivacy policy