Google launches Gemini 4 Argon with 1M token output, claims lead on DeepSWE and cybersecurity benchmarks
Google has introduced Gemini 4 Argon, a new frontier AI model designed for complex, long-horizon tasks in software engineering, enterprise knowledge work, and cybersecurity defense. The model features an industry-leading 1 million token output limit, up from 64,000 tokens in the previous generation. Google is rolling out Argon initially to trusted cyber defenders through its Fairwind Program and…
Key points
- Gemini 4 Argon launches with 1M token output limit and introductory pricing of $2/$10 per million input/output tokens.
- Google claims Argon leads on DeepSWE v1.1 (77.9%), Vals Index (68.9%), and cybersecurity benchmarks; trails on FrontierSWE v2 and OSWorld-2.0.
- Model available first to Fairwind Program cyber defenders; thousands of Googlers use it internally for coding, research, and infrastructure optimization.
The sources agree on core specifications and internal use cases but differ on benchmark comparisons and competitive positioning. The official Google blog and Digit.in emphasize Argon outperforming GPT-6 Astra on several benchmarks, while Marktechpost provides a detailed breakdown showing Argon leads on 12 of 18 benchmarks—including DeepSWE v1.1 (77.9%), Vals Index (68.9%), and AutomationBench (51.3%)—but trails GPT-6 Astra on FrontierSWE v2 (55.0% vs 65.5%) and OSWorld-2.0 (69.2% vs 72.6%), and trails Claude Opus 5.5 on Terminal-Bench 4.0. Artificial Analysis independently scores Argon 53 on its Intelligence Index, noting it generates more tokens per task than the median. Only Marktechpost reports that pricing will rise to $4 input and $20 output after the introductory period and that Logan Kilpatrick confirmed the launch pricing.
Model pages: Gemini 4 Argon → · GPT-6-Astra → · Claude Opus 5.5 →
The story so far
7 episodes →- Google launches Gemini 4 Argon with 1M token output, claims lead on DeepSWE and cybersecurity benchmarksthis story
Google introduces Gemini 4 Argon AI model, claims it outperforms GPT 6 Astra on several benchmarks
digit.in · 30 September 2026
Loading the full article…
This text was published by digit.in. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
24sources- Hacker News discussion · 43 pointsnews.ycombinator.com
- Hacker News discussion · 6 pointsnews.ycombinator.com
- Reddit discussionreddit.com
- [IT News] October 2026, Week 1 #44 | Google returns to the top 3 with Gemini 4 Argon, OpenAI blocks 'distillation' targeting reasoning processes, and President Trump renamesPress · note.com ·
- Alphabet’s New Argon Model Trails Anthropic’s Claude in Frontier AI RacePress · blockonomi.com ·
- Can Google’s Gemini 4 Argon Catch OpenAI and Anthropic?Press · coincentral.com ·
- Has AI Started to Stop Working Alone? The Next AI Emerging from OpenAI, Gemini 4, and JevPress · note.com ·
- Google unveils latest AI model, but Wall Street wants a breakout personal agentPress · CNBC Technology ·
- Argon aims to return Google to the frontierNewsletter · The Rundown AI ·
- Google Just Undercut Anthropic With Its New Gemini 4 Argon AI ModelPress · finance.yahoo.com ·
- [Explanation] Google Begins Partial Rollout of New AI ModelPress · note.com ·
- What is Google Gemini?Press · Zapier Blog ·
- Artificial Analysis says Gemini 4 Argon (high) matches GPT-6 Astra (max) on its Intelligence Index and has a 15% hallucination rate, compared with 51% for AstraPress · artificialanalysis.ai ·
- Google Unveils Gemini 4 Argon AI Model With Aggressive Pricing StrategyPress · blockonomi.com ·
- Google’s New AI Model Just Undercut OpenAI and Anthropic on PricePress · coincentral.com ·
- Google’s next big Gemini upgrade beats Claude and ChatGPT at their own gamePress · androidpolice.com ·
- Google Unveils Gemini 4 Argon, its Most Powerful AI Model YetPress · analyticsinsight.net ·
- Google Gemini 4 Argon closes the gap with OpenAI and Anthropic but doesn't take a clear leadPress · The Decoder ·
- Google DeepMind Unveils Gemini 4 Argon with 1M Output Tokens for Coding, Knowledge Work and Cyber DefensePress · MarkTechPost ·
- Gemini 4 Argon (High): Intelligence, Performance and Price AnalysisPress · artificialanalysis.ai ·
- Google announces Gemini 4 Argon AI model, but you can't use it yetPress · Ars Technica AI ·
- Introducing Gemini 4 ArgonPrimary source · blog.google ·
- Gemini 4 Argon: our next era of frontier intelligencePrimary source · Google DeepMind ·
- [AI News Analysis] Google Releases "Gemini 4 Argon" — Why Cutting-Edge AI Isn't Immediately Available to the PublicPress · note.com ·
- Google Gemini 4 Argon released: 1 million output tokens and a focus on cybersecurity, coding, and agentic capabilitiesPress · notebookcheck.net ·
- Google’s Gemini 4 Argon Closes the Pricing Triangle. The Benchmark Lead Is the Real Story.Press · finance.yahoo.com ·
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Generative AI & Models
All →- Meta releases Muse Spark Open, an 816B-parameter open-weight multimodal model with 42B active parameters and 1M-token context · 22 src
- Liquid AI releases d1-ultra decision model at $0.042 per million tokens · 1 src
- Microsoft AI chief says in-house model beats OpenAI's using quarter of compute · 1 src
- NVIDIA releases Nemotron 3.5 Super VL and Bolt edge models for enterprise agents · 8 src
- Black Forest Labs releases FLUX 3 Image Klein, a 5.6B-Parameter Open-Weights model for text-to-image generation · 3 src
Comments
via GitHub Discussions