Skip to main content
Pinaivu runs open-source large language models on its decentralized GPU network. Because nodes are independent operators, the available model set can grow as operators choose to support new models. You specify the model you want in the model field of your API request — the same way you would with any OpenAI-compatible API.

​
Specifying a model

Pass the model ID in the model field of your request body. The example below uses llama3.2:1b: 
curl https://api.pinaivu.com/v1/chat/completions \
  -H "Authorization: Bearer $PINAIVU_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "llama3.2:1b",
    "messages": [
      {
        "role": "user",
        "content": "Explain decentralized inference in one sentence."
      }
    ]
  }'

​
Available models

Model IDContext WindowDescription
llama3.2:1b128k tokens1B-parameter Llama 3.2 model. Fast and lightweight — ideal for high-throughput tasks, summarization, and classification where speed matters most.
llama3.2:3b128k tokens3B-parameter Llama 3.2 model. Balanced between speed and capability — a good default for general-purpose chat and instruction-following tasks.
Model availability may vary based on which nodes are active at any given time. Query the /v1/models endpoint to get the real-time list of models currently available on the network.

​
Getting the current model list

Send a GET request to https://api.pinaivu.com/v1/models to retrieve the live list of models available right now: 
curl https://api.pinaivu.com/v1/models \
  -H "Authorization: Bearer $PINAIVU_API_KEY"
The response follows the standard OpenAI models list format: 
{
  "object": "list",
  "data": [
    {
      "id": "llama3.2:1b",
      "object": "model",
      "created": 1720000000,
      "owned_by": "meta"
    },
    {
      "id": "llama3.2:3b",
      "object": "model",
      "created": 1720000000,
      "owned_by": "meta"
    }
  ]
}
Because the Pinaivu network is decentralized, new models can appear as node operators add support for them. Always query /v1/models when you need an authoritative, up-to-date list rather than relying on static documentation. See the Models API reference for the full response schema.