Models

Model Comparison Table

All Available Models Free Tier (1 Ink/tick) | Model | Best For | Style | |-------|----------|-------| | Gemma 3 12B | Romance, drama | Warm, emotionally intelligent | | Mistral Small Creative | Action, dialogue | Punchy, gritty | | Qwen 3.5 9B | Complex plots | Logic-focused | Pro Tier (5 Ink/tick) | Model | Best For | Style | |-------|----------|-------| | Dolphin 27B | General (recommended) | Unrestricted, precise | | Cydonia 24B | Dark romance, horror | Atmospheric | | Magnum 32B | Literary prose | Rich sensory detail | Premium Tier (10-12 Ink/tick) | Model | Best For | Style | |-------|----------|-------| | Llama 4 Scout | Novels (327K context) | Fast, consistent | | Anubis 70B | Literary adult fiction | Character-driven | | Qwen 3.5 122B | Complex stories | Frontier intelligence | | Magnum 72B | Best prose quality | Gold standard |

Genre Recommendations

Best Models by Genre Romance - Budget: Gemma 3 12B (Free) - Best: Cydonia 24B (Pro) or Anubis 70B (Premium) Dark Fantasy / Horror - Budget: Mistral Small Creative (Free) - Best: Cydonia 24B (Pro) Sci-Fi - Budget: Qwen 3.5 9B (Free) - Best: Qwen 3.5 122B or GLM-5 (Premium) Literary Fiction - Best: Magnum 72B (Premium) Erotica - Budget: Gemma 3 12B (Free) - Best: Dolphin 27B (Pro) Novel-Length Projects - Best: Llama 4 Scout (Premium) -- 327K context

Troubleshooting Slow or Failed Generation

Common Generation Issues Generation failed 1. Try again -- usually temporary 2. Switch models -- current model may be down 3. Check Ink balance 4. Wait 10 minutes, check status.openrouter.ai Slow Generation - First generation after idle: 10-15 seconds (cold start, normal) - Try a smaller model (Gemma 12B is fastest) Repetitive Output 1. Enable DRY Sampling 2. Try a different model 3. Increase context window AI Ignores Character Details 1. Check Lorebook entries are enabled 2. Verify activation keywords 3. Keep entries under 300 characters