agno-models

Configure Agno agents by parsing model strings like openai:gpt-4o, instantiating provider classes directly, or setting the AGNO_MODEL environment variable to dynamically select from 40+ providers including Anthropic, Google, Groq, and Ollama.

Switch between OpenAI and Anthropic models dynamically by changing the AGNO_MODEL environment variable, such as updating it from openai:gpt-4o to anthropic:claude-sonnet-4-5, without touching your underlying agent code.

Adjust LLM parameters like temperature, max tokens, and top_p in Agno through direct class instantiation or configuration settings to fine-tune model output generation behavior for your specific agent requirements.

Install provider-specific extras to enable full functionality for Ollama, Groq, and other LLM APIs, as the configuration requires specific dependencies for each provider's API integration to work correctly.

Agno supports 40+ language model providers including OpenAI, Anthropic, Google Gemini, Groq, and Ollama, allowing you to select between them using simple model strings or direct class imports for agent configuration.

Model string parsing in Agno requires the exact format of provider:model, such as google:gemini-2.0-flash, along with the correct provider-specific extras installed to ensure the LLM API integration functions properly.