# Gemini 4 Argon

Gemini 4 Argon is an AI model released by Google on Sep 30 2026. It has a 1M context window. Tracked results include 88.8% on LAB-Bench 2, 65.4% on Finance Agent v2 and 91.7% on LVBench.

## Facts

| Field | Value |
| --- | --- |
| Model | Gemini 4 Argon |
| Developer | Google |
| Release date | Wednesday, Sep 30 2026 |
| Licensing | Closed |
| Context window | 1M |

## API pricing

All rates in USD per 1,000,000 tokens, pay-as-you-go.

| Tier | Input | Cached input | Output |
| --- | --- | --- | --- |
| Introductory | $2.00 | $0.10 | $10.00 |

Output prices include reasoning tokens.
Introductory pricing through November 30, 2026. After that, $4 per million input tokens and $20 per million output tokens apply.
Publicly available on the paid Gemini API as gemini-4-argon since October 13, 2026. A cyber-permissive build is allowlist-only through the Fairwind Program.
Verified October 19, 2026 against the first-party source: https://ai.google.dev/gemini-api/docs/pricing

## Tracked benchmark scores

| Benchmark | Score | Source | What it measures |
| --- | --- | --- | --- |
| BullshitBench v2 | 66% | [BullshitBench](https://github.com/petergpt/bullshit-benchmark) | Given a confidently-worded but nonsensical prompt, does the AI spot that it makes no sense and push back — instead of playing along and inventing an answer? The score is how often it clearly called out the nonsense. Higher is better. |
| Gray Swan IPI (k = 15) | 0.7% | Lab | Attackers hide malicious instructions inside content the AI reads — a web page, an email, a document — and try to hijack what it does. This variant gives the attacker 15 tries and counts an attack as successful if any of them works. Lower is better. |
| CursorBench 4.0 | 52.9% | [CursorBench](https://cursor.com/cursorbench), retrieved 2026-10-22 | Cursor's own test of coding agents on ambiguous, multi-file tasks taken from real Cursor sessions — editing, refactoring, investigating a codebase, understanding what the user meant, managing jobs and following a design. Cursor runs each model at several reasoning efforts; each release here carries the score of its best listed effort. Scores aren't comparable with earlier CursorBench versions. Higher is better. |
| DeepSWE 1.1 | 77.9% | Lab | Artificial Analysis' independent test of deep, agentic software-engineering work — the AI has to plan and carry out substantial coding tasks end to end. (Version 1.1 of the test.) Higher is better. |
| FrontierSWE V2 | 55% | Lab | Engineering problems that would occupy a person for days: systems implementation, performance work, scientific computing and AI research, with up to twenty hours per task. Partial credit is awarded, because finishing one outright is still rare. Higher is better. |
| Next.js Evals | 100% | [Next.js Evals](https://nextjs.org/evals) | Vercel's open eval of how well AI coding agents build and migrate real Next.js apps — measured as the share of tasks the agent completes successfully. Higher is better. |
| Terminal-Bench 4.0 | 57.4% | Lab | Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? Version 4.0 recalibrated how much time, CPU and memory each task gets, removed eight tasks and fixed nineteen, so fewer runs fail for reasons that have nothing to do with the model. Scores are not comparable with earlier versions. Higher is better. |
| Terminal-Bench-Science 0.1 | 57.6% | Lab | The same command-line setup as Terminal-Bench, pointed at scientific work: the AI has to drive research tooling and computational workflows through to a result, rather than administer a machine. Version 0.1 is the first release of the task set, and scores run lower than on the general board. Higher is better. |
| RoboDojo Sim | 34.2% | Lab | control of a simulated two-arm robot from camera images, average success rate |
| RiemannBench | 76% | Lab | mathematical problem-solving evaluation published by Surge |
| ARC-AGI-2 | 91.25% | [BenchLM](https://benchlm.ai), retrieved 2026-10-22 | Puzzle-style tests of abstract reasoning and pattern-finding — the kind of thing people find easy but AIs often struggle with. Higher is better. |
| LAB-Bench 2 | 88.8% | Lab | Everyday tasks from a working biology lab — reading protocols, interpreting figures and sequence data, and answering the practical questions a researcher hits at the bench. Higher is better. |
| CWE-bench v1 | 68% | Lab | Collinear AI’s programmatically graded security-patch pass@1 rate; separate from exploit-generation benchmarks |
| OSWorld 2.0 | 69.2% | Lab | Can the AI actually operate a computer — clicking, typing, and using real apps — to finish tasks on its own? Version 2.0 is a harder, refreshed task set. Higher is better. |
| Agent's Last Exam (pass@1) | 39.5% | Lab | A hard set of desktop and operating-system tasks an AI agent has to finish by looking at the screen and working the machine itself. The score is the share it passes outright — partial credit does not count. Higher is better. |
| AutomationBench | 51.3% | Lab | Tests whether the AI can run real multi-step business workflows — the kind of end-to-end office processes companies want to automate — from start to finish. Higher is better. |
| Finance Agent v2 | 65.4% | Lab | Tests the AI on real financial-analysis work, like digging through reports and making sound decisions. Higher is better. |
| Harvey's Legal Agent Benchmark | 19.6% | Lab | Harvey's test of whether an AI agent can complete real legal work — drafting and reviewing documents, working with spreadsheets and presentations, and navigating files the way a lawyer's assistant would. Higher is better. |
| Chartography (no tools) | 71.6% | Lab | The same chart-centred test with no tools: the AI has to read each chart unaided. Scores run far lower than the with-tools version, so read the two as separate tests. Higher is better. |
| LVBench | 91.7% | Lab | Can the AI follow a very long video — up to an hour — and answer questions that need details from far apart in it? Higher is better. |
| GraphWalks BFS F1 (up to 128K) | 99.7% | Lab | breadth-first search F1 on 650 problems with up to 128K tokens of context |
| GraphWalks BFS (256K–1M) | 84.2% | Lab | breadth-first search F1 on the 200 GraphWalks problems with 256K to 1M tokens of context |
| threejseval | 1714 | [threejseval](https://threejseval.com/ranking) | Every model gets the same prompt — "the Eiffel Tower", "a glass fishbowl", "a robot arm picking toys into a box" — and builds a 3D scene in Three.js. Real people then see two scenes side by side, names hidden, and vote for the one they prefer. The votes become a chess-style Elo rating on threejseval.com, averaged across all the prompts. It measures whether the scene looks and moves right to a human eye, not whether the code passes a test. Higher is better. |

## About Gemini 4 Argon

Gemini 4 Argon was announced on September 30, 2026 with initial access for trusted defenders in Google's Fairwind Program. A wider rollout followed on October 13 through the paid Gemini API and selected Gemini app and Search experiences. Google positioned the model for sustained coding, research and professional workflows.

Argon's documented token limit was 1,048,576 shared across input and output. Long responses used a continuation mechanism beyond the default output allocation, so this was not simply a promise of one million output tokens in every request. Google emphasized keeping agents working through extended tool-driven tasks. Gemini 4 Carbon followed on October 28 with a further focus on sustained engineering and cybersecurity defense.

## Questions and answers

### When was Gemini 4 Argon released?

Gemini 4 Argon was released by Google on Wednesday, Sep 30 2026.

### Who made Gemini 4 Argon?

Gemini 4 Argon was built by Google. Builds the Gemini family of models through Google DeepMind. Integrates AI across Google products.

### How much does Gemini 4 Argon cost?

Gemini 4 Argon costs $2.00 per million input tokens and $10.00 per million output tokens through the Google API. Cached input is $0.10 per million tokens. Output prices include reasoning tokens. Rates are pay-as-you-go API prices verified against Google's published pricing on October 19, 2026.

### What benchmark scores did Gemini 4 Argon get?

Gemini 4 Argon reports 23 tracked benchmark scores — BullshitBench v2: 66%; Gray Swan IPI (k = 15): 0.7%; CursorBench 4.0: 52.9%; DeepSWE 1.1: 77.9%; FrontierSWE V2: 55%; Next.js Evals: 100%; Terminal-Bench 4.0: 57.4%; Terminal-Bench-Science 0.1: 57.6%; RoboDojo Sim: 34.2%; RiemannBench: 76%; ARC-AGI-2: 91.25%; LAB-Bench 2: 88.8%; CWE-bench v1: 68%; OSWorld 2.0: 69.2%; Agent's Last Exam (pass@1): 39.5%; AutomationBench: 51.3%; Finance Agent v2: 65.4%; Harvey's Legal Agent Benchmark: 19.6%; Chartography (no tools): 71.6%; LVBench: 91.7%; GraphWalks BFS F1 (up to 128K): 99.7%; GraphWalks BFS (256K–1M): 84.2%; threejseval: 1714. Tracked scores may come from lab reports or independent benchmarks; source details accompany the benchmark data. It holds the best score among all models tracked here on LAB-Bench 2, Finance Agent v2 and LVBench.

### What is the context window of Gemini 4 Argon?

Gemini 4 Argon has a context window of 1M. Providers may specify separate input and output token limits.

### Is Gemini 4 Argon open source?

No. Gemini 4 Argon is a closed model. The weights are not published — it is available only through the provider's own API, apps, or partner platforms.

### What came before and after Gemini 4 Argon?

Google's previous tracked release was Gemini 3.8 Flash Cyber on Sep 2 2026, 28 days earlier. It was followed by Gemini 4 Argon Cyber Permissive on Oct 13 2026.


---

Canonical page: https://aireleasetracker.com/model/google/gemini-4-argon
Site index: https://aireleasetracker.com/llms.txt
Source: AI Release Tracker (https://aireleasetracker.com). Most benchmark scores come from lab launch material; gathered results identify the leaderboard that published them.