Every frontier model.
One gateway.
Token Hub meters every request per token with micro-dollar precision, fails over across upstream routes automatically, and speaks the OpenAI API you already use.
Featured models
View all 22 →A creativity-tuned Claude for long-form writing, narrative synthesis, and careful editing.
The fastest Claude tier, built for high-volume production traffic and latency-sensitive products.
The frontier reasoning flagship of the Claude 5.5 family, built for long-horizon agentic work and large-document analysis.
The balanced workhorse of the Claude 5.5 family — near-frontier reasoning at a fraction of flagship pricing.
A throughput-first tier delivering frontier-adjacent quality at commodity prices.
A fast, efficient DeepSeek workhorse for high-throughput chat and coding.
Why Token Hub
One API, every model
OpenAI-compatible /v1 endpoints across the whole frontier catalog. Switch models with a single string — no SDK rewrites.
Automatic failover
Requests retry across independent upstream routes before you ever see an error. One endpoint, several nine of reliability.
Micro-dollar metering
Per-token billing with integer micro-dollar precision — no rounding drift, no surprise overages, full usage transparency.
Start in 30 seconds
https://tokenhub-staging.coality.io/v1from openai import OpenAI
client = OpenAI(
base_url="https://tokenhub-staging.coality.io/v1",
api_key="YOUR_TOKEN_HUB_API_KEY",
)
resp = client.chat.completions.create(
model="glm-5-3",
messages=[{"role": "user", "content": "Hello, Token Hub!"}],
)