AI Models Powering ChatLab
ChatLab utilizes some of the most advanced AI models available today to provide a superior chatbot experience. These models include various versions of OpenAI's GPT, Google's Gemini, and Anthropic's Claude. Each model offers unique strengths tailored to different needs and budgets.
Choosing Your AI Model
From the main ChatLab administration page, go to Chatbots, click on selected chatbot, select the Settings tab, then click Model & Advanced in the sidebar. Here you can select between various OpenAI, Google Gemini, and Anthropic Claude AI models from the Model dropdown. Each option shows its credit cost per message right in the list:
How Message Credits Work
Each model shows a credit cost per message next to its name. That number is the cost at the default 8k context size. If you increase the context window in Model & Advanced, the credit cost is multiplied:
- 8000 tokens - x1 (the cost shown next to the model)
- 16000 tokens - x2
- 32000 tokens - x4
So a 5-credit model runs at 5 credits per message at 8k, 10 credits at 16k, and 20 credits at 32k. Bigger models cost more credits per message, and a larger context multiplies that further - keep this in mind when balancing answer quality against your monthly credits.
OpenAI Models
-
GPT 5.6 Sol (15 credits per message): OpenAI's most capable model in the GPT 5.6 family. Best when answer quality matters more than cost or speed - complex reasoning, deep analysis, and demanding multi-step tasks.
-
GPT 5.6 Terra (10 credits per message): The high-capability member of the GPT 5.6 family. Strong reasoning and accuracy for demanding production chatbots, at a lower cost than Sol.
-
GPT 5.6 Luna (5 credits per message): The balanced GPT 5.6 model. Latest-generation quality at the same price as GPT 5.4 - a good default upgrade path for production chatbots that need solid contextual understanding.
-
GPT 5.5 (10 credits per message): A top-tier reasoning model. Great for highly demanding tasks - complex problem solving, advanced analysis, and tasks where accuracy matters more than speed or cost.
-
GPT 5.4 (5 credits per message): A strong reasoning model balancing quality and price. Good fit for most production chatbots that need solid contextual understanding at a moderate credit cost.
-
GPT 5.4-mini (3 credits per message): A reasoning-optimized mini model. Stronger at multi-step questions than standard Mini models, while costing less than full models. Useful sweet spot when 1-credit Mini quality is not enough.
-
GPT 5.1 (5 credits per message): A capable OpenAI model with superior reasoning and accuracy. Good choice for complex tasks requiring deep understanding.
-
GPT 5 (5 credits per message): A powerful model with excellent performance across a wide range of tasks. Great for e-commerce integrations and backend systems.
-
GPT 5 Mini (1 credit per message): A fast and cost-efficient version of GPT-5. Default choice for new bots, offering a strong balance between quality and cost.
-
GPT 4.1 (5 credits per message): An advanced model with enhanced accuracy and reasoning. Well-suited for tasks needing deeper understanding and reliability.
-
GPT 4.1 Mini (1 credit per message): A lightweight, faster version of GPT-4.1, providing a solid balance between speed and intelligence for everyday tasks.
-
GPT 4o (5 credits per message): OpenAI's multimodal model offering top-tier performance at a competitive cost. Note: does not support dynamic suggested questions or products view features.
-
GPT 4o Mini (1 credit per message): Fast and affordable OpenAI model, perfect for general-purpose interactions. Note: does not support dynamic suggested questions or products view features.
Google Gemini Models
-
Gemini 3 Pro (10 credits per message): Google's most intelligent model with breakthrough reasoning capabilities. Excels at complex problem-solving, multimodal understanding (text, images, audio, video, PDFs), and coding tasks.
-
Gemini 3.5 Flash (5 credits per message): Google's newest fast model. Brings quality close to the Pro tier while keeping Flash-level speed - a strong mid-price option in the Gemini family.
-
Gemini 3 Flash (3 credits per message): A fast model from the Gemini 3 family. Offers significantly improved reasoning and quality over earlier Flash generations while maintaining high speed.
-
Gemini 3.1 Flash-Lite (1 credit per message): The lightest and cheapest Gemini option - a 1-credit alternative to OpenAI's Mini models for high-volume FAQ chatbots. Note: does not support dynamic suggested questions or products view features.
-
Gemini 2.5 Pro (10 credits per message): A powerful alternative to GPT models, performs well in coding and structured data tasks.
Anthropic Claude Models
-
Claude Opus 4.8 (15 credits per message): Anthropic's newest most capable model, designed for complex reasoning, long-form analysis, and nuanced multi-step tasks. Best when answer quality matters more than cost or speed.
-
Claude Opus 4.7 (15 credits per message): The previous Opus generation. Still excellent for complex reasoning and long-form analysis if you prefer a longer-proven model.
-
Claude Sonnet 5 (10 credits per message): Anthropic's latest balanced flagship. Strong reasoning and writing quality at a mid-tier price - a good default for production chatbots that need consistently high answer quality.
-
Claude Sonnet 4.6 (10 credits per message): The previous Sonnet generation, with the same balanced quality-to-price profile.
-
Claude Haiku 4.5 (5 credits per message): Anthropic's fast and cost-efficient model. Solid quality at a price comparable to OpenAI's Full models, with Anthropic's signature writing style. Good alternative when you want a non-OpenAI, non-Google option.
Note: GPT-4o family models (GPT 4o, GPT 4o Mini) and Gemini 3.1 Flash-Lite do not support the dynamic suggested questions and products view features. If you have these features enabled and switch to one of these models, they will be automatically disabled and a notice appears in the settings.
Which Model Should I Choose?
- Best value for most use cases: GPT 5 Mini (1 credit) - fast, affordable, default for new bots
- Best quality (OpenAI): GPT 5.6 Sol (15 credits) - the most capable OpenAI option, or GPT 5.6 Terra / GPT 5.5 (10 credits) for top-tier reasoning at a lower cost
- Best quality (Google): Gemini 3 Pro (10 credits) - top-tier reasoning and multimodal understanding
- Best quality (Anthropic): Claude Opus 4.8 or Claude Opus 4.7 (15 credits) - complex reasoning, long-form analysis, nuanced multi-step tasks
- Balanced latest generation: GPT 5.6 Luna (5 credits), Gemini 3.5 Flash (5 credits), or Claude Sonnet 5 (10 credits)
- Reasoning sweet spot: GPT 5.4-mini or Gemini 3 Flash (3 credits) - reasoning-optimized, better than 1-credit Minis on multi-step questions
- Budget-friendly options: GPT 4o Mini, GPT 4.1 Mini, GPT 5 Mini, or Gemini 3.1 Flash-Lite (1 credit) - fast and efficient for everyday conversational tasks
- E-commerce and integrations: GPT 5.6 Luna, GPT 5.4, GPT 5, or GPT 4.1 (5 credits) - reliable for API functions and product data
To understand how credits are counted across your whole account, see What is Message Credit.