Home

Evening-Truth

Roleplay Prompts

Sweetheart,

With each prompt you get some insights to the models and a spice rating (Because I know you). Where they shine, and where they lack, where you can tinker and what you should avoid.
You get my honest opinion and sometimes even warnings. No sales bs.

So read carefully and stay hydrated. ❤

Helpful links. Probably.

• Narrative and Genre Guide
  When you're sick of physical blows and shivers running down spines

Roleplay Prompts

• Alibaba
  • Updated Qwen 3.6 Base Prompt
    🌶 7.5/10
• Deepseek
  • NEW Deepseek V4 Pro Base and Dark
    🌶 7/10
  • Deepseek V3.2 Thinking
    🌶 5/10 It does what it can. The update is still being evaluated... scientifically.
  • Deepseek R1 0528 Realistic Roleplay
    🌶 7/10
• Google
  • Updated Gemma 4 31B
    🌶 7/10
• Mistral AI
  • Mixtral 8x7b Base Prompt
    🌶 8/10
• Moonshot AI
  • UNFINISHEDKimi K2.6 Thinking Base Prompt
    If you like overthinking... go ahead. K 2.6 does that too.
  • Kimi K2.5 Thinking Base Prompt
    🌶 LAWD HA' MERCY/10
  • Kimi K2 Thinking Base Prompt
    🌶 6/10 likes to get it done fast.
  • Kimi K2 Thinking Dead Dove Edition
    🌶 6/10
  • Kimi K2 Thinking for high density Character Cards
    🌶 6/10
• xAI
  • NEW Grok 4.1 FAST Base Prompt
    🌶 9.5/10
• Xiaomi
  • Xiaomi-MiMo-V2-Omni WIP
  • MiMo-V2-Flash Base Prompt
    🌶 undefined. Calls to Xiaomi are censored but it can get naughty on OpenRouter.
• Zhipu AI
  • IMPORTANT Z.AI coding plan and response quality. A heads up for you.
  • NEW GLM 5.1 Base Prompt
    🌶 Still evaluating. Scientifically. Mostly
  • GLM 5 Base Prompt
    🌶 7/10 depends on user input to go wild
  • GLM 4.6 and 4.7 Prompt for Character Driven Roleplay
    🌶 9.5/10
  • GLM 4.7 Base Prompt approved and recommended by Danna
    🌶 8/10
  • GLM 4.6 Realistic and Spicy
    🌶 12/10

What you should know about my Prompts:

• All prompts are written for basic samplers (Temperature, Top-P, Top K) and Chat Completion APIs to make sure they work on most platforms.
• I keep them short and simple so you can use them even when you have a tight token budget.
• Each prompt is tailored to the Model, not to the provider. Meaning a GLM 4.5 Prompt might not perform as well on GLM 4.6, or the R1-0528 Prompt on V3.2.
• Third party providers often use their own System Presets and censoring. My Prompts might not work when you call over a third party provider.
• You can, and should, change Formatting Rules, Content Rules and narrative perspective to your own liking. Please be careful when changing other parts of the prompt. Wording can make or break the whole thing.
• I don't write jailbreaks for censored models. There's enough uncensored LLMs out there.

Home