Gemini 4 Argon API
google/gemini-4-argonGemini 4 Argon turns code, documents, images, audio, and video into detailed written answers, with three thinking levels and a 1,048,576-token window. Set an output budget of 32K or more so deep reasoning still leaves room for the full reply.
Gemini 4 Argon
Google · chat-completions
Quick start
Send a POST request to /v1/chat/completions with google/gemini-4-argon. Choose JSON or SSE with stream, and receive generated content and usage.
- Endpoint
- POST /v1/chat/completions
- Model ID
- google/gemini-4-argon
- Protocol
- Chat Completions
Request fields
| Field | Requirement | Description |
|---|---|---|
| model | Required | google/gemini-4-argon |
| messages | Required | A non-empty ordered array with system, user, or assistant roles and text content. |
| max_tokens | Optional | An integer from 1 to 65,536 that caps thinking and answer tokens together; the playground defaults to 32,768. |
| stream | Optional | Defaults to false for complete JSON; true returns SSE events. The playground streams replies. |
Response and usage
Successful responses include generated content and usage. Billing settles from actual input, output, and cache tokens.
Streaming
Set stream: true to read SSE. Finalize usage accounting from the terminal usage event.
curl --request POST \
--url "https://api.vidgo.ai/v1/chat/completions" \
--header "Authorization: Bearer $VIDGO_API_KEY" \
--header 'Content-Type: application/json' \
--data '{
"model": "google/gemini-4-argon",
"messages": [
{
"role": "system",
"content": "You are a helpful assistant."
},
{
"role": "user",
"content": "Review this retry loop, identify failure modes, and propose tests for the fix."
}
],
"max_tokens": 32768,
"stream": false
}'Keep API keys on the server. Never expose them in browser code or public repositories.
Continue with
Related Models
Gemini 4 Argon API FAQ
What is the Gemini 4 Argon API?
Gemini 4 Argon is a Google DeepMind model that turns code, documents, images, audio, and video into detailed text answers, and it is the first flagship of the Gemini 4 generation. Google built it for long-horizon software engineering, such as large code migrations, and for enterprise document work. It reasons inside a 1,048,576-token budget with low, medium, or high thinking. Use the API tab for Chat Completions and Gemini Native examples, or test a text conversation in the Playground.
How does the Gemini 4 Argon context window work?
Gemini 4 Argon has one 1,048,576-token budget shared by your input, its thinking, and its output. A 100K-token prompt leaves about 948K tokens for reasoning and the reply. Each response returns up to 65,536 tokens, so split very long deliverables across turns.
How do I set the Gemini 4 Argon thinking level?
In a Gemini Native request, set generationConfig.thinkingConfig.thinkingLevel to low, medium, or high; medium is the default. Use low for quick extraction and high for migrations or multi-step analysis. Because thinking counts toward maxOutputTokens, raise the budget to 32,768 or more when you choose high.
Can Gemini 4 Argon analyze video and PDF files?
Yes. Gemini Native requests accept image, audio, video, and PDF data through contents[].parts[].inlineData, combined with text instructions in the same turn. The model replies in text, for example with a summary, a table of findings, or answers that cite timestamps or page sections.
What does Gemini 4 Argon cost on Vidgo?
Input costs $1.60 (320 credits) and output costs $8.00 (1600 credits) per 1M tokens. Thinking tokens are billed as output, so a high thinking level adds to output usage. Google lists $2 input and $10 output per 1M tokens as its introductory price through 30 November 2026.
Why does Gemini 4 Argon take longer on hard prompts?
At higher thinking levels, Gemini 4 Argon can reason for several minutes before it answers. Google recommends client timeouts of 15 minutes or more for demanding requests. Streaming with streamGenerateContent or stream: true shows the answer as it arrives.
How does Gemini 4 Argon compare with Gemini 3.8 Flash?
Gemini 4 Argon targets depth: long migrations, cross-document analysis, and high thinking budgets. Gemini 3.8 Flash costs less per token and suits faster everyday coding and chat. Both accept Chat Completions and Gemini Native requests, so you can test the same prompt on each.