Docspace

Search documentation…

Pricing

Docspace

Search documentation…

Pricing

Docspace

Docs

Core Concepts

Models

Learn how AI models power every request in NovaAI. This guide explains what models are, how to choose the right one for your application, and the key characteristics that influence performance, cost, speed, and output quality.

What Are Models?

A model is the intelligence behind every NovaAI request. It processes your input, understands context, and generates an appropriate response based on its capabilities.

Different models are optimized for different workloads. Some prioritize reasoning and accuracy, while others focus on speed, cost efficiency, or multimodal understanding.

Rather than using a single model for every task, NovaAI provides a family of specialized models designed for specific use cases.

Choosing the right model can significantly improve response quality, latency, and overall application cost.

NovaAI Model Family

Model	Best For	Speed	Cost
`nova-chat-4`	General conversation and reasoning	Medium	Medium
`nova-chat-4-lite`	Fast responses and lightweight assistants	Fast	Low
`nova-reason-1`	Complex reasoning and problem solving	Medium	High
`nova-vision-1`	Image understanding and visual analysis	Medium	Medium
`nova-image-1`	Text-to-image generation	Medium	Medium
`nova-embed-1`	Embeddings and semantic search	Fast	Low
`nova-audio-1`	Speech recognition and synthesis	Fast	Medium

Choosing the Right Model

Selecting the appropriate model depends on your application's requirements.

If you need...	Recommended Model
Conversational AI	`nova-chat-4`
High-speed responses	`nova-chat-4-lite`
Advanced reasoning	`nova-reason-1`
Image analysis	`nova-vision-1`
Image generation	`nova-image-1`
Semantic search	`nova-embed-1`
Speech applications	`nova-audio-1`

Model Capabilities

Every model is designed with a specific purpose.

Capability	Chat	Vision	Image	Embedding	Audio
Text Generation	✓	—	—	—	—
Reasoning	✓	✓	—	—	—
Image Understanding	—	✓	—	—	—
Image Creation	—	—	✓	—	—
Vector Embeddings	—	—	—	✓	—
Speech Processing	—	—	—	—	✓

Model Selection Workflow

Rendering diagram…

Using a Model

Specify the model when creating a request.

1const response = await client.chat.create({
2  model: "nova-chat-4",
3  messages: [
4    {
5      role: "user",
6      content: "Explain quantum computing in simple terms."
7    }
8  ]
9});

Changing the model property allows you to switch between different AI capabilities without modifying the rest of your request.

Model Properties

Property	Type	Description
`id`	String	Unique model identifier.
`name`	String	Human-readable model name.
`context_window`	Integer	Maximum input context supported.
`supports_streaming`	Boolean	Indicates whether streaming is available.
`supports_vision`	Boolean	Indicates image understanding support.
`supports_tools`	Boolean	Indicates function calling support.
`status`	String	Current availability of the model.

Best Practices

Select the smallest model capable of completing your task.
Use reasoning models only when advanced problem-solving is required.
Choose lightweight models for real-time applications.
Keep model selection consistent across similar workflows.
Benchmark multiple models before deploying to production.

Start development with a faster model to reduce iteration time, then evaluate larger models if your application requires higher-quality reasoning or more accurate responses.

Frequently Asked Questions

Can I switch models without changing my code?

Yes. In most cases, updating the model parameter is sufficient, as NovaAI maintains a consistent request format across supported models.

Which model should I use for chat applications?

For most conversational experiences, nova-chat-4 provides the best balance between reasoning quality, speed, and cost.

Are all models multimodal?

No. Each model is optimized for a specific capability. For example, nova-vision-1 processes images, while nova-embed-1 generates vector embeddings.

Can I switch models without changing my code?

Yes. In most cases, updating the model parameter is sufficient, as NovaAI maintains a consistent request format across supported models.

Which model should I use for chat applications?

For most conversational experiences, nova-chat-4 provides the best balance between reasoning quality, speed, and cost.

Are all models multimodal?

No. Each model is optimized for a specific capability. For example, nova-vision-1 processes images, while nova-embed-1 generates vector embeddings.

Next Step

Now that you understand how NovaAI models work, continue to Prompts to learn how well-structured instructions influence model behavior and response quality.

Models

What Are Models?

NovaAI Model Family

Choosing the Right Model

Model Capabilities

Model Selection Workflow

Using a Model

Model Properties

Best Practices

Frequently Asked Questions

Next Step

Related Documentaion

Prompts

Tokens

Chat Completion