> ## Documentation Index
> Fetch the complete documentation index at: https://chatbase.co/docs/llms.txt
> Use this file to discover all available pages before exploring further.
# Models Comparison
> Compare the recommended AI models by message credit cost, and pick the one that fits your workload
## Overview
Every model available to your AI agent consumes a different number of **message credits** per message, and offers different strengths in return. This page covers the recommended models at each credit tier, so you can balance capability against cost.
See what you get at each credit tier, and choose the cheapest model that still handles your hardest conversations.
This page covers the **recommended** models. For the complete list of every available model and its credit cost, see [Build → AI Model](/docs/user-guides/chatbot/build#ai-model).
## At a glance
| Model | Credits / message | Provider | Best for |
| ------------------------------------------- | ----------------- | --------- | -------------------------------------------------------------- |
| [Auto](#1-credit-per-message) | 1 | Chatbase | Mixed workloads where you'd rather not pick a model at all |
| [GPT-6 Luna](#1-credit-per-message) | 1 | OpenAI | High-volume, low-latency tasks |
| [GPT-5.6 Terra](#2-credits-per-message) | 2 | OpenAI | The majority of production workflows |
| [Gemini 3.6 Flash](#2-credits-per-message) | 2 | Google | Solid reasoning at low latency |
| [Claude 4.6 Sonnet](#3-credits-per-message) | 3 | Anthropic | Everyday workflows needing a balance of intelligence and speed |
| [GPT-5.5](#4-credits-per-message) | 4 | OpenAI | Highly demanding enterprise workflows |
| [Claude 4.6 Opus](#5-credits-per-message) | 5 | Anthropic | Complex multi-step tasks and agentic workflows |
## Recommended models by credit cost
### 1 credit per message
**Auto**: Engineered by Chatbase for peak performance, speed, and efficiency. It adapts to every conversation, giving instant answers for everyday questions and frontier-grade intelligence for demanding ones, with multilingual and image understanding built in. One choice that handles every turn end to end, tools included.
**GPT-6 Luna**: The fastest and most cost-efficient tier of OpenAI's GPT-6 family, optimized for high-volume, low-latency tasks while retaining the core capabilities of the 6 series.
### 2 credits per message
**GPT-5.6 Terra**: The balanced tier of OpenAI's GPT-5.6 family, pairing strong reasoning and context retention with everyday speed and cost efficiency. Ideal for the majority of production AI workflows.
**Gemini 3.6 Flash**: A Google reasoning model that delivers fast, high-quality responses. It handles complex questions, multi-step instructions, and image understanding with ease, making it a strong fit for AI agents that need solid reasoning at low latency.
### 3 credits per message
**Claude 4.6 Sonnet**: An Anthropic model in the Sonnet series that builds on 4.5 with stronger reasoning, improved instruction following, and better performance on complex multi-step tasks. It offers an excellent balance of intelligence and speed for everyday workflows.
### 4 credits per message
**GPT-5.5**: One of OpenAI's most capable models, delivering notably stronger reasoning, improved context retention, and better performance across highly demanding tasks. It delivers high accuracy and efficiency for enterprise workflows.
### 5 credits per message
**Claude 4.6 Opus**: A powerful model in the Opus series that builds on 4.5 with stronger reasoning, improved instruction following, and better performance on complex multi-step tasks. It excels at demanding coding challenges, agentic workflows, and sophisticated problem-solving while managing context efficiently.
## How to choose
Rather than starting from the most capable model, start from the cheapest tier that handles your hardest conversations. Weigh these dimensions against your own traffic:
| Dimension | What to ask |
| ----------------------------- | ------------------------------------------------------------------------------------------------------------------------------------- |
| **Technical capability** | Can it answer your most complex questions correctly? |
| **Empathy and communication** | Does its tone suit your customers and your brand? |
| **Speed** | Is the response time acceptable on your busiest channel? |
| **Taking actions** | Does it reliably trigger your [actions](/docs/user-guides/chatbot/actions/actions-overview): escalating, booking, changing a subscription? |
| **Multi-step tasks** | Does it hold context and follow your [procedures](/docs/user-guides/chatbot/procedures/procedures-overview) through to the end? |
Don't judge these from the descriptions alone. Open **Build > Instructions** and click **Compare** to run the same message against several models side by side. See [Compare area](/docs/user-guides/chatbot/build#compare-area).
Test your hardest real questions against two or three tiers before committing. If the cheaper model handles them, the extra credits buy you nothing.
## Next steps
See every available model and its credit cost, and set the one your AI agent uses.
Run the same message against several models at once.
Instructions and data sources affect answer quality more than the model does.
Track how many message credits your AI agents are consuming.