How to Use Ox Alpha (Free AI Model)

What is Ox Alpha?

Ox Alpha is ZAI's GLM-5.3 Flash model, positioned for coding, multimodal work, and agentic workflows. Public descriptions associate it with a 1M-token context window and up to 128K output. Depending on the provider route, the same model may be addressed as stealth/ox-alpha on Tokenra or glm-5.3-flash on a ZAI-compatible route.

These model characteristics are useful planning signals, not a guarantee that every provider route exposes every modality, limit, or parameter. For the current integration boundary, see the Ox Alpha API reference. For a disclosure-first explanation of the model identity and evidence, read What is Ox Alpha?.

Use Ox Alpha via API

The documented Tokenra route uses an OpenAI-compatible Chat Completions request. Use stealth/ox-alpha for the Tokenra route; if your ZAI-compatible provider exposes the same model under its native identifier, use glm-5.3-flash instead. The API is the better fit for a product, repeatable evaluation, or agent workflow because your server can own credentials, validation, retries, permissions, and audit logs.

Get your API key

1. Create or access a Tokenra account.
2. Create an API key in the provider dashboard and verify that the route is available to your account.
3. Store the key in a server-side environment variable such as TOKENRA_API_KEY.
4. Never embed the key in browser JavaScript, static HTML, a client bundle, or a public repository.

Example: Python

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://tokenra.io/v1",
    api_key=os.environ["TOKENRA_API_KEY"],
)

response = client.chat.completions.create(
    model="stealth/ox-alpha",
    messages=[{
        "role": "user",
        "content": "Explain this function in two bullets.",
    }],
)

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

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://tokenra.io/v1",
    api_key=os.environ["TOKENRA_API_KEY"],
)

response = client.chat.completions.create(
    model="glm-5.3-flash",
    messages=[{
        "role": "user",
        "content": "Explain this function in two bullets.",
    }],
)

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

Example: curl

curl https://tokenra.io/v1/chat/completions \
  -H "Authorization: Bearer $TOKENRA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "stealth/ox-alpha",
    "messages": [
      {"role": "user", "content": "Summarize this request."}
    ]
  }'

curl https://tokenra.io/v1/chat/completions \
  -H "Authorization: Bearer $TOKENRA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-5.3-flash",
    "messages": [
      {"role": "user", "content": "Summarize this request."}
    ]
  }'

Start with the smallest request that proves authentication, routing, and response parsing. Then test prompt size, output limits, tool schemas, multimodal content, timeouts, and error responses against the live route. The complete field reference is in the Ox Alpha API documentation.

Free limits & pricing

Ox Alpha is currently listed as free by the provider. Free availability does not necessarily mean unlimited requests or unrestricted context. Rate limits, route availability, usable context, maximum output, and any overage terms can change independently.

For production planning, budget from observed completed requests and the provider's current commercial terms. Do not infer Tokenra pricing from a third-party benchmark's cost field. See the Ox Alpha benchmarks page for the distinction between a public benchmark record and a verified provider route.

FAQ