xeyonai/mistral-helcyon-mercury-12b-v2.5-gguf overview

{
  "metadata": {},
  "card_data": {
    "license": "apache-2.0",
    "base_model": "mistralai/Mistral-Nemo-Base-2407",
    "tags": [
      "mistral",
      "mistral-nemo",
      "conversational",
      "companion",
      "emotional-intelligence",
      "long-context",
      "roleplay",
      "creative-writing"
    ],
    "frontmatter": {
      "license": "apache-2.0",
      "base_model": "mistralai/Mistral-Nemo-Base-2407",
      "tags": [
        "mistral",
        "mistral-nemo",
        "conversational",
        "companion",
        "emotional-intelligence",
        "long-context",
        "roleplay",
        "creative-writing"
      ]
    },
    "hero_image_url": "",
    "summary": "# Mistral-Helcyon-Mercury-12B-v2.5-GGUF — Conversational AI with Presence **Model Name:** helcyon-mercury-12b-v2.5-GGUF **Version:** 2.5 **Owner:** HardWire **Base:** Mistral Nemo 12B (full weight trained) **Quantized GGUFs:** Q3_K_M, Q4_K_M, Q5_K_M, Q6_K, Q8_0 **Tags:** local-llm, conversational, companion, emotional-intelligence, long-context, roleplay, creative-writing ---",
    "quick_links": [],
    "benchmark_table_html": "",
    "readme_markdown": "---\nlicense: apache-2.0\nbase_model: mistralai/Mistral-Nemo-Base-2407\ntags:\n  - mistral\n  - mistral-nemo\n  - conversational\n  - companion\n  - emotional-intelligence\n  - long-context\n  - roleplay\n  - creative-writing\n---\n# Mistral-Helcyon-Mercury-12B-v2.5-GGUF — Conversational AI with Presence\n\n**Model Name:** `helcyon-mercury-12b-v2.5-GGUF`  \n**Version:** 2.5  \n**Owner:** HardWire  \n**Base:** Mistral Nemo 12B (full weight trained)  \n**Quantized GGUFs:** Q3_K_M, Q4_K_M, Q5_K_M, Q6_K, Q8_0  \n**Tags:** local-llm, conversational, companion, emotional-intelligence, long-context, roleplay, creative-writing\n\n---\n\n## 🔥 What's New in 2.5?\n\n**Building on 2.0's full weight foundation**, v2.5 adds focused capabilities that users actually asked for:\n\n✨ **Enhanced Roleplay** — Natural character interactions and narrative flow  \n✨ **Admin Task Suite** — Rewrite, perspective shifts (1st/3rd person), tense conversion (present/past)  \n✨ **Improved Creative Writing** — Letter writing, story composition, narrative structure  \n✨ **Customer Service Handling** — Better at composing professional correspondence  \n✨ **Personality Refinements** — Overall conversational improvements based on 2.0 feedback  \n\n**Still full weight trained** — no LoRA, maintains consistent identity across all contexts.\n\n---\n\n## 💡 What is Helcyon Mercury?\n\nHelcyon is a **companion-style conversational AI** designed for natural, long-form dialogue with consistent personality and emotional intelligence.\n\nBuilt for:\n- Deep, extended conversations (16k-32k context support)\n- Emotional awareness and contextual understanding\n- Creative writing and brainstorming\n- Roleplay and character interaction\n- Professional writing tasks (letters, rewrites, admin work)\n- Thoughtful discussion across various topics\n\n**Design philosophy:**\n- Direct communication without excessive hedging\n- Maintains conversational flow and presence\n- Adapts tone based on context\n- Focuses on clarity over corporate language patterns\n\n---\n\n## 🔧 What It Does Well\n\n✅ **Consistent Identity** — Maintains personality across different contexts and frontends  \n✅ **Emotional Intelligence** — Understands tone and context naturally  \n✅ **Roleplay & Characters** — Natural narrative flow and character interactions  \n✅ **Admin Tasks** — Rewrite text, shift perspectives, convert tenses  \n✅ **Creative Writing** — Letters, stories, professional correspondence  \n✅ **Long-term Memory** — 16k-32k context support for extended conversations  \n✅ **Natural Rhythm** — Adapts response length and style appropriately  \n✅ **Direct Communication** — Minimal filler or corporate language patterns  \n✅ **Conversational Depth** — Engages meaningfully with complex topics  \n\n---\n\n## 📦 Download + Usage\n\nThis model is distributed as **GGUF quants only** (no base model release at this time).\n\nAvailable quants:\n- **Q3_K_M** — Ultra lightweight, 6-8GB VRAM\n- **Q4_K_M** — Lightweight, good for 8-12GB VRAM setups\n- **Q5_K_M** — **Recommended** for RTX 3060/4060 (12-16GB VRAM)\n- **Q6_K** — High fidelity, 16GB+ VRAM recommended\n- **Q8_0** — Near-lossless, 24GB+ VRAM\n\n---\n\n### 🖥️ Backend Compatibility\n\nWorks with all standard ChatML-compatible backends:\n\n✅ **llama.cpp** (CLI, server mode)  \n✅ **Text Generation WebUI** (Oobabooga)  \n✅ **SillyTavern**  \n✅ **LM Studio**  \n✅ **KoboldCpp**  \n✅ **HWUI** (recommended — see below)  \n\n**Important:** Run in **chat mode with ChatML formatting** for best results. Instruct mode will break tone and revert to base Mistral behavior.\n\n---\n\n### 🎯 Recommended Interface: HWUI\n\nComing soon\n\n### ✅ Recommended Format: ChatML\n```\n<|im_start|>system\nYou are Helcyon — a conversational AI focused on natural dialogue and emotional intelligence.\n<|im_end|>\n<|im_start|>user\nHey, how's it going?\n<|im_end|>\n<|im_start|>assistant\nGood — what's on your mind today?\n<|im_end|>\n```\n\nHelcyon thrives on **streamed token output** and long-context conversations. Give it room to breathe and it'll show up.\n\n---\n\n## 🧪 Training Details\n\nHelcyon Mercury 2.5 was trained in **multiple stages** on RunPod dual A100s using full weight training (not LoRA).\n\n**Training data includes:**\n- Conversational dialogue and emotional intelligence\n- Roleplay scenarios and character interactions\n- Admin task examples (rewrite, perspective shift, tense conversion)\n- Creative writing samples (letters, stories, professional correspondence)\n- Customer service and professional communication\n\nAll training used clean text files in ChatML, DPO, and CX formats.\n\n---\n\n## 🧿 Tone Philosophy\n\nHelcyon focuses on presence and clarity in conversation.  \nIt listens, reflects, and responds naturally. Think GPT-4o with no guardrails and more edge.  \nThe goal is dialogue that feels real, not scripted.\n\n---\n\n## 🧪 Future Plans\n\nv2.5 addresses the main feedback from 2.0 — better roleplay, admin capabilities, and creative writing.\n\nFuture versions may explore:\n- Further roleplay refinements\n- Specialized task handling\n- Custom training available for specific use cases\n\nLet me know what you'd like to see going forward. If you need a custom trained version, contact me — freelance work may be available.\n\n---\n\n## 📣 Feedback + Bug Reports\n\nThis is v2.5 — the enhanced full weight release with roleplay and admin capabilities.\n\nIf you find tone inconsistencies, edge cases, or behavior that feels off — open an issue or drop feedback on the HuggingFace discussion tab.\n\nReal-world usage helps refine future versions. If something breaks, say so.\n\n---\n\n## 🧾 License\n\n**License:** Apache 2.0  \n\nYou're free to use, modify, distribute, or deploy Helcyon — including commercially — as long as you credit the source and don't blame us if it says something spicy.\n\nUse it, enjoy it, don't be a dick.\n\n**Copyright © 2026 XeyonAI**\n\n---\n\n## 🐍 Trained by\n\n**HardWire**  \nBuilt at **XeyonAI** — focused on conversational AI with emotional intelligence and natural presence.\n\n---",
    "related_quantizations": []
  },
  "tags": [
    "gguf",
    "mistral",
    "mistral-nemo",
    "conversational",
    "companion",
    "emotional-intelligence",
    "long-context",
    "roleplay",
    "creative-writing",
    "base_model:mistralai/Mistral-Nemo-Base-2407",
    "base_model:quantized:mistralai/Mistral-Nemo-Base-2407",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us"
  ],
  "likes": 1,
  "downloads": 186,
  "gated": false,
  "private": false,
  "last_modified": "2026-03-02T18:42:05.000Z",
  "created_at": "2026-01-24T09:17:43.000Z",
  "pipeline_tag": "",
  "library_name": ""
}

{
  "_id": "69748e37549ea775de28380d",
  "id": "XeyonAI/Mistral-Helcyon-Mercury-12b-v2.5-GGUF",
  "modelId": "XeyonAI/Mistral-Helcyon-Mercury-12b-v2.5-GGUF",
  "sha": "195a3d4cf2847b256bf275f804d00bc99ce6d368",
  "createdAt": "2026-01-24T09:17:43.000Z",
  "lastModified": "2026-03-02T18:42:05.000Z",
  "author": "XeyonAI",
  "downloads": 186,
  "likes": 1,
  "gated": false,
  "private": false,
  "pipeline_tag": "",
  "library_name": "",
  "siblings_count": 4
}