One API. Every Model.
1,516 routable models across 104 inference providers (3,109 models and 792 catalog entries in the full registry). Images, videos, music, audio, editing, multimodal, text, semantic search, classification, and translation — all through a single endpoint.
Suggestion: Add capability filters like web_search or vision in your API payload — the router narrows to matching models automatically.
Generative AI
Representative model lines below are drawn from the live registry and change as the catalog grows.
Image Generation
Nano Banana, Flux 1 Schnell, Flux 2 Flex, Flux 2 Klein 4b, Seedream 4.0, and more — generate photorealistic, artistic, or stylized images from text prompts. Supports LoRA fine-tuning for consistent brand styles.
From $0.004 per imageVideo Generation
Veo 3, Seedance 2 Image to Video, Seedance 2 Reference to Video, V6, Q3 Turbo — create cinematic video from text or image prompts. Supports camera motion control, character consistency, and scene transitions.
From $0.05 per secondMusic & Audio
Suno V4.5 All, Suno V5, Music 2.6 for full song and instrumental generation. Nova 3, Whisper Large V3 Turbo for transcription. Speech 2.8 Hd, Speech 2.8 Turbo for natural speech synthesis.
From $0.0043 per minuteImage Editing
Background removal, inpainting, outpainting, upscaling, and style transfer across GPT Image 2 API, Fibo Edit [Erase by Text], Fibo Edit [Relight]. Edit existing images with text prompts — change objects, extend canvases, or enhance resolution.
From $0.004 per editMultimodal Generation
Vision, image-guided generation, and mixed text-and-image inputs. Combine prompts with reference images, video frames, or audio for complex creative workflows.
From $0.01 per requestText (LLM)
Claude Fable 5, Claude Opus 4.6, Claude Opus 4.7, Claude Opus 4.8, Claude Sonnet 4, and more — chat, reasoning, code generation, instruction following, and creative writing. Long-context models support windows up to 2M tokens.
From $0.10 / 1M tokensCore AI
Text Generation
Claude Fable 5, Claude Opus 4.6, Claude Opus 4.7, Claude Opus 4.8, Claude Sonnet 4, and more — chat, code, reason, create.
From $0.10 / 1M tokensClassification
Distilbert Sst 2 Int8, Resnet 50 — content moderation, sentiment analysis, categorization.
From $0.0000025 / requestSemantic Search & Translation
Bge Large En V1.5, Embeddinggemma 300m for semantic search and similarity. Indictrans2 En Indic 1B for 200+ language translation.
From $0.008 / 1M tokensAPI Features
Cost Optimization
Set optimization: "price-optimized" or a budget_dollars cap per request, and GreatRouter automatically picks the cheapest capable model. No manual price comparisons — the routing engine scores options in real time across 1516 models from 792 providers.
Every request includes the actual cost in the response metadata. Track spending by model, provider, or task type from your dashboard. Auto-recharge keeps your balance topped up so you never hit a service interruption.
Budget per Request
Pass budget_dollars and the router excludes any model whose estimated cost exceeds the cap. Works for text, image, video, and audio requests.
Price Tiers
Filter by price tier — Economy, Standard, Balanced, Premium, or Flagship — to narrow the model pool before routing. Each tier represents a pricing percentile within its category, so Balanced models sit near the average cost.
Fine-Grained Parameters
Control routing with precise parameters. Everything is optional — use as much or as little control as you need.
task_typetext_generation, text_to_image, text_to_video, music_generation, etc.content_modeauto, generate, edit — controls whether to create new content or modify existingcapabilitiesArray — vision, reasoning, function_calling, web_search, realtime, lora, batchproviderPin to a specific provider — openai, anthropic, google, meta, etc.budget_dollarsMaximum estimated cost in USD for this request. The router excludes models above the cap.maxCosteconomy, standard, balanced, premium, flagship — filters by model pricing percentile within its categoryIntent Classification
GreatRouter reads your prompt and automatically detects the intent — chat, image generation, transcription, translation, and more.
Content Mode
generate creates new content from a prompt. edit modifies existing content. auto lets the router decide based on context.
Tools and Functions
Web Search
Models with web search capability can access real-time information from the internet. Route to models like GPT-4o with search for up-to-date answers.
{"model": "router",
"capabilities": ["web_search"],
"messages": [...]} Function Calling
Send tool definitions and get structured JSON responses. Works with GPT-4o, Claude, Gemini, Mistral, and other models that support function calling.
{"model": "router",
"capabilities": ["function_calling"],
"tools": [...],
"messages": [...]} Vision
Send images alongside text prompts. Route to multimodal models that can analyze, describe, and reason about visual content.
{"model": "router",
"capabilities": ["vision"],
"messages": [
{"role": "user",
"content": [
{"type": "image_url",
"image_url": {"url": "..."}},
{"type": "text",
"text": "Describe this"}]}]} Reasoning
Route to models optimized for complex reasoning — math, logic, analysis. Includes o1, DeepSeek R1, and other chain-of-thought models.
{"model": "router",
"capabilities": ["reasoning"],
"messages": [...]}