Enterprise LLM Endpoint

​

Quick start

from openai import OpenAI

client = OpenAI(
    base_url="https://modelslab.com/api/v1/enterprise/proxy/v1",
    api_key="YOUR_ENTERPRISE_API_KEY",
)

response = client.chat.completions.create(
    model="default",
    messages=[{"role": "user", "content": "Write a haiku about GPUs."}],
)
print(response.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://modelslab.com/api/v1/enterprise/proxy/v1",
  apiKey: "YOUR_ENTERPRISE_API_KEY",
});

const response = await client.chat.completions.create({
  model: "default",
  messages: [{ role: "user", content: "Write a haiku about GPUs." }],
});
console.log(response.choices[0].message.content);

curl https://modelslab.com/api/v1/enterprise/proxy/v1/chat/completions \
  -H "Authorization: Bearer $ENTERPRISE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "default",
    "messages": [{"role": "user", "content": "Write a haiku about GPUs."}]
  }'

​

What you get

• The full request body passes through. Streaming, tools and function calling, structured output (response_format), reasoning settings, and prompt caching (cache_control) all work as the model supports them.
• Files, images and PDFs in the messages, as a URL or base64. See Files, Images and PDFs.
• Streaming with "stream": true returns Server-Sent Events as the model generates tokens.
• Usage in every response. The usage object reports prompt, completion, and cached tokens.
• A fixed price. Your enterprise plan sets the price. You do not pay per token.

​

What you cannot change

• The model. The server replaces model with your plan’s model. It also removes the fields that pick other models (models, route) and saved presets (preset).
• The endpoints. By default the endpoint accepts chat/completions, completions, and models. Other paths return 404.

​

Limits

• A request that is not streamed must finish in about 110 seconds. For long outputs, set "stream": true.
• A request can be at most 30 MB, including files sent as base64.
• A trial plan stops at its expiry date. After that, requests return 403. See Errors.

​

Endpoints