The AI Chat Plugin in SeaTable supports language models from three providers: Anthropic, OpenAI, and Mistral. You choose which model you want to use and bring your own API key. The resulting token costs are paid directly to the respective provider, not to SeaTable.

Each provider offers two models: a fast, affordable model for simple tasks and a more powerful model for demanding analyses.

Provider Model Strengths Cost tendency
Anthropic Claude Haiku Fast, efficient, good for simple queries low
Anthropic Claude Sonnet Powerful, precise for complex analyses medium
OpenAI GPT-4o mini Fast, cost-efficient, versatile low
OpenAI GPT-4o Very capable, high response quality higher
Mistral Mistral Small Fast, affordable, good baseline performance low
Mistral Mistral Large Powerful, good for complex tasks higher

To use a model in the AI Chat Plugin, you need an API key from the respective provider. You create this directly on the provider’s website:

All three providers require you to add a payment method. Costs are based on the number of tokens processed and are billed directly by the provider.

For looking up individual values, short summaries, or simple filter questions, the affordable models are well suited: Claude Haiku, GPT-4o mini, or Mistral Small. These deliver fast answers at low cost.

If you want to evaluate large datasets, identify patterns, or perform multi-step calculations, choose Claude Sonnet, GPT-4o, or Mistral Large. These models deliver more precise results for demanding tasks.

For composing emails, reports, or descriptions based on your data, Claude Sonnet or GPT-4o are recommended. Both models produce linguistically high-quality text.

All token costs are billed directly between you and the respective AI provider. SeaTable does not charge any additional fees for using the AI Chat Plugin. You can find the exact per-token pricing on the pricing page of the respective provider.