In short
- Anthropic and OpenAI.Auto picks the model for you. You can pin a Claude or GPT model.
- Auto, if you want it.Auto picks the model for you. By default it moves up a tier only for language.
- Costs shown upfront.Every model shows its typical credits per answer before you choose it.
- Tiers in plain words.Models are grouped as Fast, Standard, Premium and Ultra. Free uses Fast and Standard.
The model behind your agent shapes how well it writes and how much each answer costs. OneType uses models from two leading AI companies, Anthropic and OpenAI, and gives you a simple choice: use Auto, which picks the model for you, or pin one model for an agent.
The models on offer
Your agent answers with Claude models from Anthropic and GPT models from OpenAI. They range from fast, low-cost models for everyday questions to the largest models for the hardest ones. Every model can use tools, and the hosted Claude and GPT models can read images that visitors attach.
- Fast: quick, low-cost answers for simple questions.
- Standard: balanced quality and cost for most support.
- Premium: stronger reasoning and tone for important conversations.
- Ultra: the most capable models, for the hardest questions.
Choosing a model
Open an agent and go to its Model tab. You see a table with every model you can use: its name, tier, typical credits per answer, and what it is best for. The first row is Auto, which we recommend for most agents. Click Use on any row to switch.
If a model is not included in your plan, the row shows which plan includes it. Once Premium with your own AI key is available and you have connected a provider key, that provider's models are marked With your own key.
Credits per answer, explained
The credits shown next to each model are a typical figure, so you can compare models before you pick. Each answer is billed by the tokens it actually used, so a short answer can cost less and a long one more. Premium with your own AI key is coming soon: every answer will count as one, whatever the model.
Mix models across agents
Each agent has its own model setting. Your support agent can stay on Auto while your sales agent uses a premium model. Change it any time and it applies to the next answer.
Same rules on every model
Whichever model answers, your agent follows the same rules. It answers from your content, cites its sources, and says it does not know when your content has no answer. Changing the model changes how well it writes, not what it is allowed to say.
If a model is slow to start or unavailable, OneType moves on to another suitable model before your visitor sees an error, so answers keep coming even when a provider has a bad moment.
Where to start
For most businesses, Auto is the right start. It picks the model for you with no setup. Pin a specific model later if you want the same model behind every answer, or if one agent handles especially important conversations. You can switch back at any time, and the change applies from the next answer.
What each plan includes, and how AI credits work, is on the pricing page.