LiteLLM integration

Use Gonka24 with LiteLLM

Use Gonka24 as an OpenAI-compatible upstream in LiteLLM SDK or Proxy with tested model, base URL, key, and streaming settings.

Configuration verified August 31, 2026

Connection details

Every example on this page uses the same Gonka24 OpenAI-compatible endpoint, a dashboard API key, and a live model ID.

Base URL: https://api.gonka24.com/v1
Authentication: Bearer sk-...
Model: GET /v1/models

Use a live model ID.

The catalog changes. Call GET /v1/models with your API key and copy an id from the response instead of relying on a model name in a guide.

bash

curl https://api.gonka24.com/v1/models \
  -H "Authorization: Bearer $GONKA24_API_KEY"

Ready-to-use config

litellm_config.yaml for LiteLLM Proxy

Expose the upstream under the stable LiteLLM alias gonka24 while keeping the current Gonka24 model ID in one place.

yaml

model_list:
  - model_name: gonka24
    litellm_params:
      model: openai/<model-id-from-/v1/models>
      api_base: https://api.gonka24.com/v1
      api_key: os.environ/GONKA24_API_KEY

Direct LiteLLM Python SDK call

python

import os
from litellm import completion

response = completion(
    model="openai/<model-id-from-/v1/models>",
    api_base="https://api.gonka24.com/v1",
    api_key=os.environ["GONKA24_API_KEY"],
    messages=[{"role": "user", "content": "Hello from LiteLLM"}],
)

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

Set up LiteLLM

Export GONKA24_API_KEY in the process that runs LiteLLM.
Replace the model placeholder with an exact ID from GET /v1/models.
For Proxy, start LiteLLM with litellm --config ./litellm_config.yaml and call the proxy model alias gonka24.
For the Python SDK, run the direct completion example and inspect choices[0].message.content.

What is supported

Supported

LiteLLM SDK and Proxy

Both paths can address Gonka24 through LiteLLM's OpenAI-compatible provider adapter.

Supported

Chat Completions and streaming

Use completion(...) for buffered calls or stream=True for OpenAI-compatible streaming chunks.

Supported

Tool-call pass-through

OpenAI-style tools can pass through LiteLLM when the selected upstream model supports them.

Not included

Responses API upstream

LiteLLM Proxy may expose /responses for other backends, but Gonka24's verified upstream contract here is /v1/chat/completions.

Verify before an agent task

Fetch the model list and copy an exact model ID.
Send a short plain chat request through LiteLLM.
Confirm the request appears in your Gonka24 dashboard activity.
Only then enable tool use, file writes, or long-running agent work.

Troubleshooting

LiteLLM chooses OpenAI directly: Prefix the upstream model with openai/ and set api_base explicitly to the Gonka24 /v1 URL.

Environment variable is not resolved: Use exactly os.environ/GONKA24_API_KEY in Proxy YAML and export the variable before starting the process.

Proxy model not found: Clients of your local LiteLLM Proxy must request the alias gonka24, not the upstream model ID.

Responses request fails: Call /chat/completions or LiteLLM completion(...); do not route a Responses API request to Gonka24.

Verification sources

Checked against the product documentation below and the Gonka24 public API contract on August 31, 2026.