The model you asked for, at full precision, with automatic failover.

One OpenAI-compatible API for frontier and open models, served by several providers, billed per token in USD.

from openai import OpenAI
import os

client = OpenAI(
    base_url="https://api.thetoken.dev/v1",
    api_key=os.environ["TOKEN_API_KEY"],
)
reply = client.chat.completions.create(
    model="moonshotai/kimi-k3",
    messages=[{"role": "user", "content": "Hello"}],
)
print(reply.choices[0].message.content)

curl https://api.thetoken.dev/v1/chat/completions \
  -H "Authorization: Bearer $TOKEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "moonshotai/kimi-k3",
       "messages": [{"role": "user", "content": "Hello"}]}'

import anthropic
import os

client = anthropic.Anthropic(
    base_url="https://api.thetoken.dev",
    api_key=os.environ["TOKEN_API_KEY"],
)
message = client.messages.create(
    model="moonshotai/kimi-k3",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello"}],
)
print(message.content)

11models 4providers …request success, 30 days, of …5.5%credit purchase fee, no hidden costs

Request success loading, from TOKEN's request logs.

Three promises

Full precision

Every model is served at the maker's released precision, never a quantized copy. Each provider's stated precision is shown with the date we checked it.

How precision is verified

Automatic failover

If a provider fails, your request moves to the next one, up to three times. You see one response at the same price.

How routing works

Clear billing

One price per model, whichever provider answers. Prepaid credit in USD, top-ups charged in SAR at 3.75 SAR per USD, and every request's cost in your usage log.

See prices

Featured models

Moonshot AIKimi K3$3.14 in · $15.68 out per 1M1M context · 3 providersMXFP4 native Z.aiGLM-5.3$1.54 in · $4.84 out per 1M1M context · 3 providersFP8 native DeepSeekDeepSeek V4 Pro$1.46 in · $4.36 out per 1M1M context · 1 providerFP8 native DeepSeekDeepSeek V4.1 Flash$0.22 in · $0.66 out per 1M1M context · 1 providerFP8 native OpenAIgpt-oss-120b$0.041 in · $0.19 out per 1M128K context · 1 providerBF16 GoogleGemma 4 31B$0.16 in · $0.44 out per 1M256K context · 1 providerBF16

All 11 models

Works with

Claude Code Cursor Cline OpenCode OpenAI SDK LangChainand any OpenAI-compatible client.

How to start

Create an account. Verify your email and get $0.50 of starter credit.
Top up. From $5, charged in SAR (cards, mada, Apple Pay), with the 5.5% fee shown at checkout.
Copy your key. Point any OpenAI-compatible client at https://api.thetoken.dev/v1.

Reliability, measured

Daily request success over 30 days. How it is measured ·Status page

Start with $0.50 of free credit

One API key for every model on TOKEN.

Get API key