Gemini 4 Argon
Google's reasoning model for code migrations, research and complex workflows.
Run this model
google/gemini-4-argon
Use an endpoint supported by this model.
Get a key for this setup →const response = await fetch('https://api.minirouter.sh/v1/chat/completions', { method: 'POST', headers: { Authorization: `Bearer ${process.env.MINIROUTER_KEY}`, 'Content-Type': 'application/json', }, body: JSON.stringify({ model: 'google/gemini-4-argon', messages: [{ role: 'user', content: 'Why is the sky blue?' }], max_tokens: 1024, }),}) const data = await response.json()console.log(data.choices[0].message.content)Start with a text request and an explicit max_tokens limit. Thinking shares that budget. MiniRouter currently supports text input for this model.
Introductory provider pricingCurrent prices include MiniRouter's markup. Google announces $4 input and $20 output per million tokens after its introductory period, without specifying an end date. Reasoning counts as output; final charges follow provider-reported usage. Source ↗
Model overview
- Input / 1M
- $2.1
- All-in MiniRouter rate
- Output / 1M
- $10.5
- All-in MiniRouter rate
- Context
- 1M
- 1,000,000 tokens
- Max output
- 1M
- tokens
- DeepSWE v1.1
- 77.9%
- Google-reported
- Reasoning
- low → high
- Three effort levels
- Announced
- 30 Sept 2026
- publisher release
Built for the longer task
Bring the whole problem
Pair your instructions with code and source material.
From code changes to completed workflows.
Three different tests, each with its own scoring and setup. These publisher results do not measure MiniRouter performance.
Publisher evidence
DeepSWE v1.1
Long-horizon software engineering in the mini-swe-agent harness. Google-computed, pass@1 at highest thinking.
Published score
0 to 100%
- Gemini 4 Argon77.9%
View chart data
| Model | Score |
|---|---|
| Gemini 4 Argon | 77.9% |
Publisher evidence
Vals Index
An aggregate across finance, coding, legal and tax work, weighted by each sector's contribution to US GDP. Reported by Google from Vals evaluations.
Published score
0 to 100%
- Gemini 4 Argon68.9%
View chart data
| Model | Score |
|---|---|
| Gemini 4 Argon | 68.9% |
Publisher evidence
AutomationBench
End-to-end business workflows on Zapier's private evaluation set. Publisher-reported; this does not measure MiniRouter routes.
Published score
0 to 100%
- Gemini 4 Argon51.3%
View chart data
| Model | Score |
|---|---|
| Gemini 4 Argon | 51.3% |
Give it a problem with a finish line.
- [ 01 ]
Change a codebase
Supply the relevant files, constraints and acceptance checks. Review the diff and run your tests before shipping.
- [ 02 ]
Reason across source material
Ask for evidence tied to supplied text. Verify claims and calculations against the originals.
- [ 03 ]
Complete a multi-step task
Provide your own tools and stopping conditions. Keep checkpoints for long runs and review consequential actions.
Introductory rates. Separate from your live price.
Introductory · per 1M tokens
- Input
- $2.00
- Cached input
- $0.10
- Output
- $10.00
View chart data
| Rate | USD |
|---|---|
| Input | $2.00 |
| Cached input | $0.10 |
| Output | $10.00 |
Set a budget, then measure the result.
- [ 01 ]
Leave room for the answer
Thinking consumes output tokens. Use an explicit token limit and compare complete answers, total cost and latency.
- [ 02 ]
Keep continuation explicit
Vercel describes long-form work across multiple requests. MiniRouter does not automatically extend a request beyond its token limit.
- [ 03 ]
Validate added modalities
Start with text. MiniRouter currently supports text input for this model.
One route, standard clients.
https://api.minirouter.sh/v1google/gemini-4-argon/v1/chat/completions/v1/messages/v1/responsesConfigured providers. Priority follows the live catalog order.
| Provider | Priority | Input / 1M | Output / 1M | Parameters | Access |
|---|---|---|---|---|---|
vercel | 1 | $2.1 | $10.5 | max_tokens, temperature, stop, tools, tool_choice, reasoning, include_reasoning | MiniRouter balance |