A model affects response quality, speed, tool use, context handling and AI usage cost.
Provider and model
The builder separates provider from model so you can compare choices from OpenAI, Anthropic, Google Gemini, Moonshot and NVIDIA where compatible. Availability and pricing can change.
Recommended defaults
Gemini 3.7 Flash is the current recommended default for cost-conscious general agent work. Higher-tier models may be better for complex reasoning, writing or specialist tasks but can cost more or respond more slowly.
Changing models
Customers can request a model change through support. We check tool compatibility, pricing, context requirements and service impact before applying a controlled migration.