> ## Documentation Index
> Fetch the complete documentation index at: https://docs.mefriend.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Model Guide: Which One Should I Choose?

> A breakdown of every AI model on Mefriend and what each one is best at

## So Many Models... Which One Do I Pick?

It can be overwhelming to see so many options! Think of these models like different types of actors or writers—some are quick and witty, some are deep and emotional, and others are strictly logical geniuses.

Here is the breakdown to help you pick the best one for your story.

***

## The "House" Models (Fast & Efficient)

<Callout title="Fast & Efficient" emoji="🏠">
  These models are optimized for speed. They are great for daily use.
</Callout>

### Lite — Free

<Card title="Best for: Casual chatting, quick replies, and simple scenes" icon="bolt">
  **The Vibe:** This is our "daily driver." It's snappy, intelligent for its size, and follows instructions surprisingly well. It doesn't have a massive memory, so it might forget details from 50 messages ago, but for a quick, fun RP session, it's excellent.

  **Under the Hood:** A highly efficient modern model known for being "punchy" and smart without needing a supercomputer to run.
</Card>

### Spark — Free

<Card title="Best for: Romance, emotional scenes, and creative storytelling" icon="heart">
  **The Vibe:** Spark is a fan favorite for a reason. It is a specialized model that has been fine-tuned specifically for roleplay. While it isn't the smartest at complex math or logic, it writes prose that feels incredibly human, warm, and creative. It avoids the robotic "assistant" tone that larger models sometimes have.

  **Under the Hood:** A specialized 8B parameter model famous in the RP community for its high creative writing quality.
</Card>

### Mimo — 3 Coins

<Card title="Best for: Long-context stories where speed matters" icon="gauge-high">
  **The Vibe:** Mimo is a "speed demon" with a very long memory. It uses a special architecture that allows it to process huge amounts of text very quickly. It's a great all-rounder—smarter than Lite, faster than Titan, and capable of remembering details from much further back in your chat history.

  **Under the Hood:** A new "Mixture of Experts" model designed for high-speed processing and massive context windows (memory).
</Card>

### Titan — 4 Coins

<Card title="Best for: Complex plots, RPG mechanics, and world-building" icon="shield-halved">
  **The Vibe:** The heavyweight of our House models. If your RP involves complex magic systems, inventory tracking, or intricate political plots, use Titan. It is incredibly smart and logical. It can be a bit more "formal" in its writing style compared to Spark, but it rarely gets confused.

  **Under the Hood:** A massive-scale model that rivals the intelligence of the top premium models (like GPT-class) but at a much lower price point.
</Card>

***

## The "Premium" Models (The State of the Art)

<Callout title="Top-Tier AI" emoji="✨">
  These are the smartest AIs in the world. They run on massive supercomputers, which is why they cost more coins per message. Premium models also support **"Thinking" (Reasoning)**—they can reason step-by-step before responding, producing more thoughtful and coherent replies.
</Callout>

### GLM 4.7 — 6 Coins

<Card title="Best for: Poetic descriptions and bilingual (English/Chinese) users" icon="feather">
  **The Vibe:** A unique model with a very distinct "flavor." Users often find its prose to be more flowery and poetic than the American models. It can be very creative, though sometimes a bit unpredictable.
</Card>

### Gemini 3.0 Flash — 8 Coins

<Card title="Best for: Massive story memory and fast responses" icon="bolt-lightning">
  **The Vibe:** The king of **Context**. If you have a chat with 100,000 words of lore, Flash can read and remember almost all of it instantly. It is very fast but slightly less nuanced in writing style.
</Card>

### Gemini 3.0 Pro — 12 Coins

<Card title="Best for: Long, complex adventures with deep lore" icon="book-open">
  **The Vibe:** A creative powerhouse. It combines that massive memory with high-level reasoning. It's great for very long, complex adventures where you need the AI to remember a minor character from Chapter 1.
</Card>

### Claude 4.5 Sonnet — 15 Coins

<Card title="Best for: The perfect balance of prose, emotion, and instruction following" icon="pen-fancy">
  **The Vibe:** The most popular high-end model. It has the perfect balance of beautiful prose, high emotional intelligence, and strict instruction following. It rarely breaks character and writes very natural dialogue.
</Card>

### Claude 4.5 Opus — 25 Coins

<Card title="Best for: The absolute best writing quality available" icon="crown">
  **The Vibe:** The luxury option. Opus is the smartest, most nuanced writer available. It is slower and expensive, but it writes like a professional novelist. It picks up on subtle subtext, sarcasm, and complex emotional beats that other models miss.
</Card>

### Claude 4.6 Opus — 25 Coins

<Card title="Best for: The latest and most capable AI model" icon="crown">
  **The Vibe:** The newest addition to the Opus family. Building on everything that made 4.5 Opus great, 4.6 pushes the boundaries even further with improved reasoning, richer prose, and even better emotional nuance.
</Card>

***

## Why Do Premium Models Cost More?

You might wonder why **Claude Opus** costs 25 coins while **Spark** is free.

It comes down to **"Brain Size" (Parameters)** and **Compute Cost**:

<CardGroup cols={2}>
  <Card title="Spark" icon="laptop">
    Like a talented short-story writer working on a laptop. It's efficient and cheap to run.
  </Card>

  <Card title="Claude Opus" icon="microchip">
    Like a room full of PhD professors working on a supercomputer. Every message requires a massive amount of electricity and computing power to process the billions of connections needed to generate that high-quality, nuanced response.
  </Card>
</CardGroup>

***

## Quick Reference: What Should I Use?

| I'm looking for...            | Best Choice             | Cost | Why pay more?                                                                                         |
| ----------------------------- | ----------------------- | :--: | ----------------------------------------------------------------------------------------------------- |
| **Unlimited Fun**             | **Spark**               | Free | Best to start. Great for casual romance & daily chat without limits.                                  |
| **Smarter RPG Logic**         | **Titan**               |   4  | Smarter than Spark. Perfect for games with stats, inventory, or complex rules.                        |
| **Infinite Memory**           | **Gemini 3.0 Flash**    |   8  | Reads everything. Can remember details from 100,000 words ago instantly.                              |
| **Complex Plot Twists**       | **Gemini 3.0 Pro**      |  12  | Upgraded Gemini Flash. Same massive memory plus deep reasoning for complex worlds.                    |
| **Deep Emotional Connection** | **Claude 4.5 Sonnet**   |  15  | Feels alive. The industry standard for beautiful prose and human-like empathy.                        |
| **The Perfect Storyteller**   | **Claude 4.5/4.6 Opus** |  25  | Ultimate Luxury. Writes like a professional novelist. Picks up on subtle subtext & sarcasm perfectly. |

<Tip>
  Not sure where to start? We recommend **Titan** for the best everyday experience — it's smart, reliable, and handles complex stories with ease. And don't forget to use your coins to try premium models like **Claude 4.6 Opus** or **Gemini 3.0 Pro** — once you experience the difference, you won't want to go back!
</Tip>