> For the complete documentation index, see [llms.txt](https://powerfan-1.gitbook.io/redchat.app/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://powerfan-1.gitbook.io/redchat.app/guides/choosing-an-ai-model.md). # Choosing an AI Model The very same character can feel completely different depending on which AI model you chat with. One model writes like a calm novelist; another fires back quick and light. There's no "right answer" here — just pick the brain that suits your taste. Spend five minutes with this guide and you'll have a feel for **what a model actually is**, **when to reach for which one**, and **what it costs**. We'll skip the jargon. > **If you only remember one thing:** **The Default model is unlimited for everyone.** Free plan, weekly limit used up, 0 RP — it doesn't matter. The conversation never stops. Every other model just nibbles a little at your "weekly premium limit" and your RP. *** ## What's a model? **Think of a model as the "brain" you lend to your character.** The character's personality, voice, and settings stay exactly the same. But the thing actually *performing* that character is the AI model. So when you swap brains on the same character, the emotional range, reasoning, and response speed can shift dramatically. * **A lighter brain:** Fast and easy-going. Perfect for everyday chat. * **A heavier brain (premium):** Thinks more deeply, remembers long context better, and paints emotions with finer detail. The tradeoff is that it uses more resources. *** ## ⚡ Two things to know before you start picking **1. Paid models skip the "cooldown."** Premium models you run on RP always respond fast — no matter how much you use them, there's no slow stretch in between. **2. The free Default model can slow down a bit under heavy use.** Lean on it hard and the Default model may briefly drop in speed (a cooldown). But **your conversation never gets cut off** — replies just arrive a touch slower. > **Tip:** Tap the **heart (♡)** next to a model you like to add it to your favorites. They gather at the top of the picker under **Favorite Models**, so they're quick to grab next time. *** ## How do I read the cost (RP)? The number next to each model is that model's **starting fee (base RP)**. For example, Opus 4.8 starts **from 720 RP**, and GLM 4.7 Flash starts **from 15 RP**. That number is a *starting point* — it **ticks up a little when you set memory deeper, write longer replies, or use a character with a large profile.** Don't worry — **the app shows you the estimated cost before you send.** [Red Points & Costs](/redchat.app/guides/red-points-and-costs.md) *** ## The model lineup Organized by brand. Each model is introduced by its **personality**, so go with whichever brain calls to you. ### 🟠 Claude — masters of prose Models with top-tier expressiveness and a gift for language. | Model | Starting RP | One-liner | | ----------------- | ----------- | ---------------------------------------------- | | **Opus 4.8** | 720 RP | Master of complex narratives and deep emotions | | **Sonnet 4.6** 🔥 | 430 RP | Balance for immersive and fast-paced roleplay | | **Haiku 4.5** | 150 RP | Light, witty, snappy banter | ### ✦ Gemini — massive context Remembers vast context and carries long worlds forward. | Model | Starting RP | One-liner | | --------------------- | ----------- | -------------------------------------------- | | **Gemini 3.1 Pro** | 300 RP | Massive context for infinite world-building | | **Gemini 2.5 Pro** 🔥 | 200 RP | Steady and logical storyteller for long arcs | | **Gemini 3.5 Flash** | 220 RP | Lightning-fast for dynamic action | | **Gemini 3 Flash** | 120 RP | Smooth, uninterrupted, efficient | | **Gemini 2.5 Flash** | 75 RP | Snappy and reliable for quick daily chats | ### ∞ GPT | Model | Starting RP | One-liner | | --------- | ----------- | ---------------------------------------------- | | **GPT-5** | 200 RP | New frontier of human-like creative expression | ### GLM — great value + cultural nuance Affordable yet sharp, and a current favorite in the roleplay community. | Model | Starting RP | One-liner | | ----------------- | ----------- | ------------------------------------------------- | | **GLM 5.2** | 190 RP | Cultural nuance and bilingual logic | | **GLM 4.6** | 95 RP | High-fidelity reasoning for sophisticated stories | | **GLM 4.5 Air** | 30 RP | Efficient everyday partner | | **GLM 4.7 Flash** | 15 RP | Ultra-fast for seamless roleplay | ### 🔷 DeepSeek — smart and economical | Model | Starting RP | One-liner | | --------------------- | ----------- | ----------------------------------------------- | | **DeepSeek V4 Pro** | 90 RP | Elite intelligence for intricate and deep plots | | **DeepSeek V3.2** | 50 RP | Consistent personality with clear reasoning | | **DeepSeek V4 Flash** | 20 RP | High-speed efficiency for non-stop action | ### Standard — default & freedom | Model | Starting RP | One-liner | | ----------------- | ------------ | -------------------------------------------- | | **Hermes 3** | 145 RP | Emotional depth and artistic flair | | **Default Model** | **Free** | Reliable and free base for standard roleplay | | **Custom Proxy** | Your own key | Ultimate freedom with your own API key | > **What's Custom Proxy?** It's the option to plug in an API key you've issued yourself. Since it runs on your own key, no redchat RP is deducted. (For advanced users.) *** ## Which one should I pick? Recommendations by situation. If you're new, start with the ones in **bold**. * **I just want to hang out and chat** → **Default (free)**, or Gemini 2.5 Flash, GLM 4.7 Flash * **Immersive drama and emotional arcs** → **Sonnet 4.6**, Hermes 3, Gemini 2.5 Pro * **A defining scene, nailed at top quality** → **Opus 4.8**, GPT-5 * **Complex plots and sprawling worlds** → Gemini 3.1 Pro, DeepSeek V4 Pro * **Smart on a budget** → GLM 4.6, DeepSeek V3.2 > **Tip:** The priciest model isn't always the right call. Enjoy a lighter model day to day, and only step up to premium when "this scene has to be perfect" — your weekly limit will stretch much further. *** ## How to switch models 1. Tap the **model icon in the bottom-right** of the chat screen to open the **Select AI Model** window. 2. Pick the model you want from the by-brand list. (Tap the heart ♡ to add it to favorites.) 3. Your very next reply uses the new model. **You can change it back anytime.** Switching models mid-conversation keeps your character settings fully intact. *** ## How do costs and usage work? In one line: > **Default model = free and unlimited** (may slow down briefly under heavy use). Every other premium model draws on your **weekly premium limit** first, and once that's spent, it continues on **RP**. What if both your weekly limit and your RP run dry? You can **keep chatting anytime on the free Default model.** There's no hard lock. [Red Points & Costs](/redchat.app/guides/red-points-and-costs.md) *** ## Frequently asked questions **Q. Is a more expensive model always better?** No. A lighter model is plenty for everyday chat. Saving premium for the moments that matter is the most economical approach. **Q. If my RP hits 0, can I no longer use premium models?** Premium models may be limited until your RP refills. But you can **always keep chatting on the Default model (free).** **Q. If I switch models, do my conversation so far and my character settings disappear?** No. Only the brain changes — your character and conversation context stay exactly as they were. **Q. Does Custom Proxy cost RP?** No. Since it runs on your own API key, no redchat RP is deducted. *** **In one line:** A model is the "brain" you lend to your character. Default is free and unlimited (may occasionally slow down), while premium models start from the RP shown next to them and scale up with your settings. Paid models are always fast with no cooldown, and even with no RP, your conversation continues on Default. **Read next:** [Red Points & Costs](/redchat.app/guides/red-points-and-costs.md) · [Response Length](/redchat.app/guides/response-length.md) · [Longer Memory](/redchat.app/guides/longer-memory.md) · [Plans & RP Usage](/redchat.app/faq/faq-plans-rp.md) Spot something missing or wrong in this doc? Let us know at — it goes a long way toward making this guide better. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://powerfan-1.gitbook.io/redchat.app/guides/choosing-an-ai-model.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.