Supported Models

The InteractiveAI Router provides access to over 200 language models across 50+ providers through a single, unified API endpoint. Rather than maintaining a static list that risks becoming outdated, the platform offers two always-current ways to explore the full model catalog: a programmatic API endpoint and an interactive UI page.

Model Identifier Format

Every model in the Router follows a consistent naming convention:

{provider}/{model-name}

For example:

• anthropic/claude-sonnet-4 — Anthropic's Claude Sonnet 4

• openai/gpt-4o — OpenAI's GPT-4o

• google/gemini-2.5-flash — Google's Gemini 2.5 Flash

• deepseek/deepseek-v3.2-speciale — DeepSeek V3.2 Speciale

• mistralai/mistral-large — Mistral's Large model

Use this identifier as the model parameter in your API requests. When the model parameter is omitted, the Router uses the default configured for your project.

Browsing Models via the Platform UI

Navigate to Orchestration → LLMs in the sidebar to access the Router LLMs page:

https://app.interactive.ai/project/{your-project-id}/router-llms

The model catalog displays a searchable, sortable table where each row represents an available model. The table includes the following information for every model:

Column	Description
Model Name	Human-readable name (e.g., "Anthropic: Claude 3.5 Sonnet")
Model ID	The identifier used in API calls (e.g., interactive/anthropic/claude-sonnet-4)
Provider	The upstream provider (Anthropic, OpenAI, Google, DeepSeek, etc.)
Best for	Brief description of recommended use cases and model strengths
Capabilities	Supported modalities such as Text, Vision, Tools, Files, and Video
Context	Maximum context window size in tokens (e.g., 200k, 1M)
Cost	Number of pricing rules configured. Click to see the full pricing breakdown

Click any model row to open its detail view, which displays complete pricing per million tokens, the match pattern used for trace identification, and a filterable history of all generations that used that model.

Use the search bar and column sorting to quickly filter models by provider, capability, or context window size. This is the fastest way to compare models for a specific use case.

Retrieving Models via API

For programmatic access, query the models endpoint to retrieve the full, up-to-date catalog:

GET https://api.interactive.ai/api/v1/models

The response returns a JSON array where each model object includes:

{
  "id": "interactive/anthropic/claude-sonnet-4",
  "object": "model",
  "created": 1770311934,
  "provider": "Anthropic",
  "marketing_name": "Anthropic: Claude Sonnet 4",
  "description": "Claude Sonnet 4 is Anthropic's...",
  "prices": {
    "input": "0.000003",
    "output": "0.000015",
    "output_reasoning": "0.000015"
  }
}

Field	Description
id	The model identifier to use in the model parameter of API calls
provider	The upstream provider name
marketing_name	Human-readable display name
description	Detailed description of the model's architecture and strengths
prices	Token pricing broken down by usage type (input, output, reasoning)

This endpoint requires no authentication and returns the complete catalog in a single request.

Supported Providers

The Router integrates with a wide range of providers, including but not limited to:

Provider	Example Models	Capabilities
Anthropic	Claude Sonnet 4, Claude Haiku	Text, Vision
OpenAI	GPT-4o, GPT-4o Mini, o1	Text, Vision, Tools
Google	Gemini 2.5 Flash, Gemini 2.5 Pro	Text, Vision
DeepSeek	DeepSeek V3.2, DeepSeek R1	Text
Mistral	Mistral Large, Mistral Small	Text, Tools
Meta (via partners)	Llama 3.1, Llama 3.3	Text
AI21	Jamba Large 1.7, Jamba Mini 1.7	Text
Aion Labs	Aion-1.0, Aion-1.0-Mini	Text
AllenAI	OLMo 3, Molmo2	Text, Vision

New providers and models are added regularly.

Choosing the Right Model

When selecting a model, consider these factors:

Task complexity vs. cost: Larger models like Claude Sonnet 4 or GPT-4o deliver stronger reasoning and instruction-following but cost more per token. For simpler tasks like classification or extraction, smaller models such as GPT-4o Mini or Mistral Small offer significant cost savings with adequate performance.

Context window requirements: If your application processes long documents or maintains extended conversation histories, prioritize models with larger context windows. Models range from 4K to over 1M tokens depending on the provider.

Capability requirements: Not all models support every modality. If your application requires vision (image understanding), tool use (function calling), or file processing, filter models by the Capabilities column in the Router LLMs page to ensure compatibility.

Latency sensitivity: Smaller models generally respond faster. For real-time applications where response speed matters, benchmark latency using the model's generation history available in each model's detail view on the platform.