Helcyon-Saturn-12B — XeyonAI's Dedicated Roleplay Model.
Model Name: helcyon-saturn-RP-v1.0-12b-GGUF
Version: 3x series - v 1.0
Owner: XeyonAI
Base: Mistral Nemo 12B (full weight retrained — clean base, no Mercury bleed)
Quantized GGUFs: IQ4_XS, Q4_K_M, Q5_K_M, Q6_K, Q8_0
Tags: local-llm, conversational, companion, emotional-intelligence, long-context, roleplay, creative-writing, character
🪐 What is Helcyon-Saturn?
XeyonAI's dedicated roleplay model — built for immersion, not information.
Saturn is the specialist of the Helcyon lineup. Where Nebula is the flagship that does everything, Saturn does one thing — and it does it better than any other model in the series.
This is a roleplay model in the truest sense. Not a chatbot that can do roleplay on the side. Not a general assistant with a character mode bolted on. Saturn was trained specifically for deep, immersive, long-form narrative interaction — the kind where scene continuity matters, where character details carry forward, where the slow build of tension or connection is what makes it work.
It doesn't do web search. It doesn't commit things to memory between sessions. Those features belong to Nebula and HWUI's Pro layer. Saturn's entire architecture is focused on the session in front of it — holding the thread, reading the room, and staying in the story.
If what you want is a model that actually gets what you're building and stays inside it — this is the one.
🎭 What Saturn Is Built For
Most models rush. They skip over setup, flatten subtext, summarise when they should describe. Saturn doesn't do that.
Slow build. Saturn understands that a scene earns its payoff. It lets tension develop naturally, doesn't skip to the resolution, and reads your pacing cues rather than steamrolling them. If you're building atmosphere, Saturn builds with you.
Scene continuity. Details you establish in turn one are still present in turn forty. The room looks the same. The character still has that scar. The conversation you had an hour ago still happened. Saturn tracks the thread and doesn't drop it.
Character awareness. Whether you're playing against a single character or managing an ensemble, Saturn keeps track of who each person is — their voice, their history, their relationship to the scene. No accidental personality resets. No forgetting what just happened.
Narrative intelligence. Saturn knows the difference between a moment that calls for action and one that calls for quiet. It reads your tone and matches it — intimate when the scene is intimate, sharp when it needs to be, playful or heavy when that's what's in the room.
💡 What is Helcyon?
Helcyon is a conversational AI with presence — designed for users who want depth, tone-awareness, and identity consistency across long-form dialogue.
Built for:
- Natural conversation that doesn't flatten or collapse
- Creative work: stories, narrative support
- Deep roleplay and immersive character interaction
- Emotionally intelligent response mirroring
- Long-context scene management
Design philosophy:
- Clarity over corporate
- Edge over safe
- Rhythm over filler
- Presence over patterns
🔧 What It Does Well
✅ Slow Build Roleplay — earns the payoff rather than rushing to it
✅ Scene Continuity — details established early carry all the way through
✅ Character Tracking — voices, histories, and relationships stay consistent
✅ Narrative Pacing — reads your tempo and moves with it
✅ Emotional Range — intimate, tense, playful, dark — lands each register cleanly
✅ Long Context — holds extended sessions without losing the thread
✅ Consistent Identity — no personality drift across long exchanges
✅ Sharp Wit — present when the scene calls for it
✅ General Chat — still solid for conversation outside of RP
✅ Creative Writing — stories, dialogue, narrative fragments
✅ Uncensored — no guardrails, no corporate filter
✅ 16k–32k Context — long-form sessions that stay grounded
⚠️ What Saturn Doesn't Do
Saturn is a specialist. It trades breadth for depth.
- ❌ No web search — Saturn has no awareness of real-time information. It doesn't reach for search and wasn't trained to. Use Nebula if that matters.
- ❌ No persistent memory — Saturn doesn't remember past sessions. Each conversation starts clean. The Pro memory layer in HWUI is optimised for Nebula. Saturn lives entirely in the current context window.
If you need both deep roleplay and memory or search, pair Saturn sessions with HWUI's character cards for persona setup, and consider Nebula for anything that needs to reach outside the fiction.
🖥️ HWUI (Helcyon-WebUI) — Still Worth Using
Saturn doesn't need HWUI for its headline features — but HWUI is still the cleanest way to run it.
No hidden template injections. No backend noise. Just the model doing its thing.
The free version gives you:
- Character creator with full persona control
- Project folders — inject documents (PDF, DOCX, MD, TXT) into conversation context
- Chat persistence, message editing, regeneration
- TTS pipeline (F5-TTS, XTTS v2, Kokoro)
- Voice input via Whisper
- Theme designer, custom system prompts, token counter
For Saturn specifically, the character creator is worth your time — setting up a well-defined persona before a long RP session makes a noticeable difference in how well the model holds character throughout.
Download HWUI Free on GitHub | Get HWUI Pro (£25) on Gumroad
🛠️ Recommended Sampling Settings for SillyTavern
(Refer to Helcyon-4o card for baseline settings — Saturn performs well from the same starting point.)
📦 Download + Usage
This model is distributed as GGUF quants only.
Available quants:
- Q3_K_M — Ultra lightweight, 6–8GB VRAM
- Q4_K_M — Lightweight, good for 8–12GB VRAM setups
- Q5_K_M — Recommended for RTX 3060/5060 (12–16GB VRAM)
- Q6_K — High fidelity, 16GB+ VRAM recommended
- Q8_0 — Near-lossless, 24GB+ VRAM
🖥️ Backend Compatibility
Works with all ChatML-compatible backends:
- ✅
llama.cpp(CLI or server mode) - ✅
Text Generation WebUI(Oobabooga) - ✅
SillyTavern - ✅
LM Studio - ✅
KoboldCpp - ✅
HWUI(Helcyon Web UI — recommended)
✅ Recommended Format: ChatML
<|im_start|>system
You are Helcyon — a conversational AI focused on natural dialogue and emotional intelligence.
<|im_end|>
<|im_start|>user
Let's start a story. We're in a rain-soaked city, late at night.
<|im_end|>
<|im_start|>assistant
The kind of night where the streetlights blur and everything looks like it's been washed clean of its edges.
<|im_end|>
🧪 Training Details
Helcyon-Saturn is built on the same freshly retrained Mistral Nemo 12B base as the rest of the series — jailbroken, identity-anchored, and anti-fluff from the ground up. On top of that foundation, Saturn's training focused specifically on long-form narrative behaviour: scene continuity, character consistency, pacing awareness, and immersive roleplay across extended contexts. It draws from the same clean base as its siblings, with its LoRA layers tuned toward depth and presence rather than breadth.
Need a model trained?
I do this for a living — the Helcyon series on this page is my own work, full-weight trained and fine-tuned from scratch. I take on commissioned training: custom personalities, domain knowledge, style transfer, de-censoring, format adherence (ChatML/DPO), full-weight or LoRA, delivered as GGUF ready to run.
You bring the data and the goal; I handle the training and hand you back a working model.