Odem is the luminous counterpart to Atem. Where Atem breathes darkness, Odem is the breath of light: radiant, epic, heroic. Built for backlit silhouettes, golden hours, and characters who look like the last thing standing between the world and the dark. The goal is output that feels like a key frame, not a screenshot.

What this model is good at

• Backlighting, rim light, god rays, light breaking through mist, trees, or doorways

• Heroic and sacred compositions: holy warriors, final stands, the morning after the battle

• Luminous hair and skin. White and silver hair especially catches light beautifully

• Warm color grading (gold, amber, sunset orange) with clean, bright highlights

• Wide, epic scenery with a sense of scale and atmosphere

What this model isn’t good at

• Not built around explicit content. It can produce it, but I didn’t tune it for that.

• Not a horror or true-darkness model. Odem pulls toward light and will fight you on pitch-black scenes. If you want menace and deep shadow, use Atem.

• Strict monochrome palettes. Odem likes to add warm accents and may drift golden under strong light. Use color negatives if you need tight control.

Recommended settings

• Sampler: [Euler Negative Dy, Restart, DPM++ 2M Karras] (Don’t use Hiresfix for Euler Negative, unless you know what you are doing.)

• CFG: [4–6]

• Steps: [30 for Euler Negative Dy]

• Resolution: Standard SDXL resolutions. [832x1216, 1024x1024, or 1216x832]

Prompt guidance

• Light is your main lever. backlighting, rim light, god rays, light rays, sunrise and golden hour do more for Odem than any quality token.

• Give the light a source. “Light from inside the gate” or “sunbeam through the trees” beats a generic dramatic lighting. Odem builds the whole composition around where the light comes from.

• Let the character look into the light. looking up toward the source turns a portrait into a moment.

• Recommended negative: lowres, worst quality, low quality, bad anatomy, bad proportions, signature

Merge composition
Built on Illustrious XL 2.0 base, merged with 2.5D style models, some lighting LoRA. Mixing

License and use
Feel free to use outputs commercially. Merging into your own work is welcome; credits are appreciated but not required.
No training data restrictions beyond what the base license carries.
If you build something interesting with this, I’d love to see it. Tag me on Twitter (@TridIsntAName) or drop a link in my DMs!