API reference

Models

List every model you can call, with its context window, privacy lane and price.

http

GET https://www.synaps.chat/api/v1/models

This endpoint is public and needs no key. The list refreshes about every 10 minutes from the routing provider, so new models appear without any change on your side.

Example

bash

curl https://www.synaps.chat/api/v1/models

The values below are illustrative. Call the endpoint for live prices.

json

{
  "object": "list",
  "data": [
    {
      "id": "deepseek/deepseek-chat",
      "object": "model",
      "owned_by": "DeepSeek",
      "name": "DeepSeek V3",
      "context_length": 163840,
      "privacy_lane": "private",
      "input_modalities": ["text"],
      "output_modalities": ["text"],
      "pricing": {
        "prompt_per_million_usd": 0.36,
        "completion_per_million_usd": 1.06
      }
    }
  ]
}

Fields

idstring

Pass this exact value as model in a chat completion.

owned_bystring

The company that makes the model.

namestring

A readable name.

context_lengthinteger

Maximum tokens of input plus output the model accepts.

input_modalitiesstring[]

What the model can read: text, plus image for vision models and file for native PDF input. PDFs work on every model; see Chat completions.

output_modalitiesstring[]

What the model can produce. Models with image can generate images.

privacy_lanestring

free, private or provider. See Privacy lanes.

pricing.prompt_per_million_usdnumber

Price per million input tokens, including the 20% Synaps fee.

pricing.completion_per_million_usdnumber

Price per million output tokens, including the fee.

Which models are listed

Synaps lists text models from Anthropic, OpenAI, Google, DeepSeek, Qwen, Moonshot, Z.ai, xAI, Meta and Mistral, newest first, plus their free variants (IDs ending in :free). The full table with list prices is on the Models page.

Choosing a model

You need	Look for
Lowest cost, high volume	DeepSeek, Qwen and small Gemini or GPT variants
Hard reasoning and code	Top Claude, GPT and Gemini models
Very long documents	Models with a context length of 1M tokens
Prompts not stored or trained on	Any model with the `private` lane
Testing for free	IDs ending in `:free`

Fetching the list in code

python

import requests
models = requests.get("https://www.synaps.chat/api/v1/models").json()["data"]
cheap = sorted(models, key=lambda m: m["pricing"]["completion_per_million_usd"])[:5]
for m in cheap:
    print(m["id"], m["pricing"]["completion_per_million_usd"])