KeyoAPI
Gemini 4 API on KeyoAPI
KeyoAPI currently lists two Gemini 4 chat IDs: gemini-4-argon and gemini-4-barium. This family guide connects their model pages, prices and OpenAI-compatible request format. Choose one of the exact IDs in your request; the family label Gemini 4 is not itself a callable model ID.
Gemini 4 models and token prices
Both IDs advertise the OpenAI-compatible chat endpoint in the public catalog. On 2026-10-13, each listed $0.7353 per million input tokens and $3.6765 per million output tokens in the default group. These are KeyoAPI sell rates; confirm current prices and your group multiplier in the pricing page before budgeting.
| Model ID | Input / 1M tokens | Output / 1M tokens | Model guide |
|---|---|---|---|
| gemini-4-argon | $0.7353 | $3.6765 | the gemini-4-argon model page |
| gemini-4-barium | $0.7353 | $3.6765 | the gemini-4-barium model page |
Call Gemini 4 Argon
Set your OpenAI client's base URL to https://www.keyoapi.xyz/v1 and use a KeyoAPI key. The example sends a plain text chat request to Argon. Replace the model value with gemini-4-barium to evaluate Barium with the same messages.
curl https://www.keyoapi.xyz/v1/chat/completions \
-H "Authorization: Bearer $KEYO_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"gemini-4-argon","messages":[{"role":"user","content":"Turn this brief into an implementation checklist: import records, validate required fields, and report rejected rows."}],"max_tokens":500}'Check the route before relying on advanced features
An OpenAI-compatible chat request does not establish that every Google-native feature is forwarded. Confirm supported content types, context limits, tool calling, streaming and optional parameters for the specific model and channel you use. Start with a small known input, validate the response shape, then increase the workload.
At equal listed token rates, compare Argon and Barium by accepted results and actual output length. Use a fixed evaluation set for your code, document or assistant workflow and log response usage. Do not infer a performance ranking from the two names. More Gemini catalog options are on the Gemini API pricing page, and both individual guides appear in the model guides page.
FAQ
Can I send model=gemini-4?
Use gemini-4-argon or gemini-4-barium. The checked public catalog does not list a bare gemini-4 ID.
Are Argon and Barium priced differently?
They had equal default-group input and output rates in the checked catalog. Confirm live prices on the pricing page, because rates and group multipliers can change.
Is this Google's native Gemini API?
The example uses KeyoAPI's OpenAI-compatible POST /v1/chat/completions route. Consult the endpoint reference for any native routes or advanced features rather than assuming identical Google-console behavior.