Model marketplace: capabilities and API pricing
The marketplace lists models available on Cloudigo, covering text, image and video models, searchable by type and name. Each model has a detail page with an introduction, context length, calling methods and unit prices; sign in to try it online. Once chosen, copy the model ID and call it through the gateway with a Cloudigo API key.
1. Compare model info and prices
Each provider keeps its own pricing page, docs and console. Comparing context length, unit price and calling methods means checking them one by one, and billing items and units differ. Selection and cost estimation take extra time.
With the Cloudigo marketplace you compare parameters, prices and calling methods in one page, then call the model directly.
2. Model coverage
Text models
Covers OpenAI’s GPT series, Anthropic’s Claude series, plus Qwen, DeepSeek, Kimi and GLM, for chat, coding assistance, content creation and long-text processing.
Image models
Image generation models are available; see the marketplace list for specific models and parameters.
Video models
Covers ByteDance’s Seedance series for generating video from text or a reference image.
3. Parameters and selection
The list supports filtering and search by model type and name. Narrow it down, then check the following on the detail page:
- Model type
- Narrow by purpose first, then judge fit from the model introduction.
- Context length
- Check this first for long documents, long conversations or large code files.
- Max output
- Look at this when a single run must produce long code, a report or an article.
- Price
- Unit prices for the model; see “Pricing and billing” below for how they are calculated.
- Calling methods
- Request formats the model supports, which determine what you need to write.
- Introduction
- What the model is designed for and what it is good at.
4. Pricing and billing
Model APIs need no subscription and are billed by actual token usage. The marketplace shows unit prices in USD per million tokens.
To estimate cost, multiply input and output token counts by their unit prices and add them up. Input that hits the cache uses the cache-read price; cache writes use the cache-write price. Models that support cache writes use a single cache-write tier regardless of cache duration. See each model’s detail page for exact prices.
Some models use tiered pricing; see the model card for tiers and prices.
Usage and cost from calls are shown in Token Usage. When the LLM account balance is insufficient, new calls are rejected; see the model gateway for top-up options.
5. Calling methods
Supported calling methods differ by model; follow the format and examples on each detail page. The Claude series, for example, supports Chat Completions and messages; GPT-5.4 and GPT-5.5 use Responses rather than Chat Completions.
Video models are called asynchronously: submit a generation task, then query task status and results. Video generation is billed by actual token usage; status and result queries are not charged separately.
6. Start calling
Filter and confirm
Filter or search in the marketplace, then confirm parameters, price and calling method on the detail page.
Copy the model ID
Copy the model ID once confirmed.
Create an API key
Create a key in API key management and call any available model through the gateway — no provider keys needed.
Configure your app
Set the base URL https://cloudigo.ai/api/v1 and the API key in your app, then send requests using that model’s calling method.
7. Frequently asked
What is the difference between context length and max output?
Context length is the total token limit per request; longer input must be split. Max output is the token limit for generated content per request, determining how long a single piece of code, report or article can be. They are independent limits — a long-context model does not necessarily produce long output in one go.
Why do some models not show max output?
Some models do not provide a max-output parameter, so the detail page omits it. Check the model’s calling documentation for its single-run output length.
How do I know if a model supports prompt caching?
Check the price items on the detail page. Models with caching list cache-write and cache-read prices separately; without those items, caching is not billed.
Choose a model and integrate
Confirm the model’s parameters and price, create an API key, and call it through the gateway.