Production inferencefor every model

Get your API keyRead quickstart

OpenAI-compatible API for Kimi, MiniMax, Z.AI GLM, GPT, Claude, DeepSeek, and open-weight models. Token pricing, per-key budgets, and usage analytics.

2API keys on Freemint scoped keys per account
$0.15/MFrom inputgpt-4o-mini on Free
1Endpointlitellm.tensorloop.tech/v1

Capabilities

Everything you need to ship inference

OpenAI-compatible APIUse the official OpenAI SDK—set base_url to litellm.tensorloop.tech/v1 and swap your API key.
Streaming responsesSSE chat completions on every plan. Stream tokens to your app without a separate integration path.
Tool callingFunction and tool schemas work on Free and Pro. Build agents without changing request shape.
Per-key dashboardsBudget, RPM, and token usage per key—so production, staging, and experiments stay legible.
Transparent token pricingPer-model $/M rates with no hidden platform fee in v1. Pay only for tokens you use.
Open models, lower costLlama 3.1 70B at $0.88/M input and output—competitive open-weight pricing alongside frontier models.

How it works

Three steps to your first completion

01

Create an account

Free tier includes gpt-4o-mini and 2 API keys. No credit card required.

02

Mint a scoped key

Set budget, RPM, and allowed models per key. Kimi and MiniMax unlock on Pro.

03

Call the API

POST to litellm.tensorloop.tech/v1/chat/completions. Swap models with one parameter change.

Supported models

Frontier and open-weight models

Kimi, MiniMax, Z.AI GLM, GPT, Claude, DeepSeek, and more through one OpenAI-compatible endpoint.

See more models
FrontierKimi K2.5$0.60/M input
CodeMiniMax M2.5$0.30/M input

Developer experience

One endpoint. Any model.

request.py
import requests

response = requests.post(
    "https://api.tensorloop.dev/v1/chat/completions",
    headers={"Authorization": "Bearer tl_your_key"},
    json={
        "model": "llama-3.1-70b",
        "messages": [
            {"role": "user", "content": "Explain transformers"}
        ],
        "max_tokens": 512
    }
)

print(response.json()["choices"][0]["message"]["content"])

Why TensorLoop

$0Platform fee in v1

Passthrough upstream pricing on every model

$5Per-key budget on Free

30-day spend cap per key · 30 RPM

5 minTo first token

Quickstart in the docs

Get started

Start building with TensorLoop

Free tier includes gpt-4o-mini, 2 API keys, and $5 budget per key. Upgrade to Pro for Kimi, MiniMax, and the full catalog.

Free

$0

Try gpt-4o-mini with scoped keys and usage caps.

  • ·2 API keys
  • ·$5 budget per key / 30d
  • ·30 requests per minute
  • ·gpt-4o-mini
  • ·Streaming & tool calling

Pro

Usage-based

Full catalog including Kimi K2.5 and MiniMax M2.5.

  • ·10 API keys
  • ·$100 budget per key / 30d
  • ·300 requests per minute
  • ·All available models
  • ·Vision on supported models
View all model pricing
Create free accountSign in