agao LogoDocs
AI Chat

Model Select

Choose and switch between different AI models for optimal performance and cost efficiency.

The Model Select feature lets you choose from various AI models optimized for different tasks and budgets. Click the model dropdown in the top navigation to see all available options. Each model displays smartness and price scores on a 1-10 scale, plus information about image support and token limits. Hover over any model to see detailed specifications before selecting.

Note: Available models depend on your specific agao instance configuration and your user group permissions. The models you see may vary based on your organization's setup, licensing, and the access rights assigned to your user group. If you need access to additional models not currently available in your dropdown, please contact your system administrator.

Choose higher smartness scores for complex reasoning tasks, consider price scores for budget optimization, and select image-capable models when working with visual content. Your selection applies immediately to new conversations and persists throughout your session.

Model Selection Interface

Available Model Providers

ProviderModels AvailableAccess Method
OpenAIGPT-5, GPT-4.1, GPT-4.1-mini, GPT-4o, GPT-4o-minivia Microsoft Azure
MetaLlama 4 Scout, Llama 4 Maverick,Llama 3.1 405B, Llama 3.3 70Bvia Microsoft Azure / via Ionos
MistralMistral-Large, Ministral 8Bvia Mistral
OpenAI Open Sourcegpt-oss-120bMicrosoft Azure / via Ionos
DeepSeekDeepSeek R1, DeepSeek V3via Microsoft Azure
xAIGrok 3, Grok 3 Minivia Microsoft Azure
MicrosoftPhi-4, Phi-4-minivia Microsoft Azure / via Ionos

Custom Providers

agao supports integration of additional model providers beyond the default offerings. Other model providers can be easily integrated through our service team for enterprise clients requiring specific AI capabilities. Additionally, any OpenAI API-compatible providers can be configured directly by system administrators through the admin interface, allowing organizations to quickly add new model endpoints without requiring custom development work.