Gemini 4 Argon
Gemini 4 Argon is Google’s flagship reasoning model for complex engineering, scientific analysis, knowledge work and autonomous multi-step agents, with multimodal input and long-context reasoning.
Examples
Input
GraphWalks 256k–1M: answer graph-traversal questions over an extremely long context.
Output
Google reports an 84.2% score for Gemini 4 Argon on this benchmark. The result describes the vendor’s evaluation setup and does not guarantee the same outcome for every prompt.
Long-context example from Google’s published model card; this is a vendor benchmark summary, not a run performed by this site.
Official specifications
Google lists the stable model code as gemini-4-argon. It uses a sparse mixture-of-experts transformer architecture and has a March 2026 knowledge cutoff.
Limits
- 1,048,576 unified context tokens
- 65,536 full-speed output tokens
- 1,048,576 maximum long-decoding output
- March 2026 knowledge cutoff
Tools and modalities
The API accepts text, image, video, audio and PDF input and returns text. Documented tools include code execution, file search, function calling, Google Search grounding, structured output, URL context and preview computer use.
Pricing and availability
Google opened Gemini 4 Argon to developers and enterprises worldwide through the Gemini API (AI Studio and the Gemini platform) on October 13, 2026. Google AI Pro and Ultra subscribers get it in the Gemini app and in AI Mode, and Ultra subscribers in Antigravity. The introductory API price is $2 per million input tokens and $10 per million output tokens, with cached input at a 95 percent discount; from November 30, 2026 the regular $4 / $20 rates apply. These are Google's list prices. What you pay in this workspace is always set by the server-side quote and the task ledger.
Google API pricing (introductory)
- $2 input / 1M tokens introductory, before Nov 30, 2026
- $10 output / 1M tokens introductory, before Nov 30, 2026
- $4 / $20 regular rates from Nov 30, 2026
- 95% cached input discount off the input price
Benchmark comparison
Google’s October 2026 model card compares Argon with GPT-6 Astra, Claude Fable 5.1 and Claude Opus 5.5. Argon leads some knowledge-work, legal, long-context and multimodal rows, while rivals lead several coding, science and computer-use rows.
| Benchmark | Gemini 4 Argon | GPT-6 Astra | Claude Fable 5.1 | Claude Opus 5.5 |
|---|---|---|---|---|
| Vals Index | 68.9% | GPT-6 Astra 63.1% | Claude Fable 5.1 65.8% | Claude Opus 5.5 67.0% |
| DeepSWE v1.1 | 77.9% | GPT-6 Astra 74.1% | Claude Fable 5.1 67.4% | Claude Opus 5.5 74.2% |
| FrontierSWE v2 | 55.0% | GPT-6 Astra 65.5% | Claude Fable 5.1 56.3% | Claude Opus 5.5 62.3% |
| Terminal-bench 4.0 | 57.4% | GPT-6 Astra 58.2% | Claude Fable 5.1 57.9% | Claude Opus 5.5 66.4% |
| GraphWalks 256k–1M | 84.2% | GPT-6 Astra 71.8% | Claude Fable 5.1 65.0% | Claude Opus 5.5 66.8% |
| LVBench | 91.7% | GPT-6 Astra 87.5% | Claude Fable 5.1 79.7% | Claude Opus 5.5 83.7% |
Artificial Analysis v4.3.2 screenshots supplied by the site owner. This is a third-party snapshot, not a benchmark run by this site. artificialanalysis.ai ↗
Limitations and responsible use
Google notes that Argon can hallucinate and may sometimes be slow or time out. Verify high-impact facts, code changes and citations, and follow the applicable acceptable-use policy.
Source: https://deepmind.google/models/model-cards/gemini-4-argon/
Start a real Argon task
The model selector is populated from this site's live catalog. Choose Google Gemini 4 Argon, review the server quote, submit the task and use the ledger as the final billing record.
Questions
Is Gemini 4 Argon officially released?
Yes. Google published the model card on October 13, 2026 and lists gemini-4-argon as the stable Gemini API model code.
Can I use Gemini 4 Argon on this site?
Yes. this site's live catalog returns gemini-4-argon-chat as available, callable and priced. The selector and quote still come from the runtime catalog.
What is the context limit?
Google documents a unified input-and-output context window of 1,048,576 tokens, with 65,536 tokens at full-speed synchronous output and up to 1,048,576 through long decoding.
Does it support images and PDFs?
Yes. Google documents text, images, video, audio and PDF as inputs, with text output. The controls shown in this workspace depend on the live catalog.
Is this a Google website?
No. This is an independent multi-model workspace and is not affiliated with, endorsed by or operated by Google.
How much does the Gemini 4 Argon API cost?
Google's introductory API price is $2 per million input tokens and $10 per million output tokens, with cached input at a 95 percent discount; regular rates of $4 / $20 apply from November 30, 2026. Charges in this workspace are set by the server-side quote and settled in the task ledger.
Explore next
Gemini 4 Argon AI chat
Chat with Gemini 4 Argon through this site's live catalog, with multimodal input, required reasoning and a server quote before every request.
Gemini 3.8 Flash
Chat with Gemini 3.8 Flash, served here today. Specs, published benchmarks, pricing and prompt patterns, plus how it differs from Gemini 4 Argon.
Gemini 4 Argon is officially released: what changed and how to use it
Gemini 4 Argon is officially released. Learn what Google confirmed, the model’s strongest use cases, important limits and how to start a real task here.
Gemini 4 Argon parameters and benchmarks: an evidence-led comparison
A sourced Gemini 4 Argon review covering its 1M-token context, multimodal inputs, official benchmark scores, trade-offs and comparisons with frontier models.