--- title: Gemini 4 Argon lands everywhere, while Anthropic builds... description: Google's Gemini 4 Argon went globally available today, landing in Copilot, Antigravity, and a half-dozen IDEs within hours of launch. Anthropic revealed it had... canonical: https://daily.dev/posts/gemini-4-argon-lands-everywhere-while-anthropic-builds-its-own-git-to-survive-claude-s-output-toun9wwtd twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Gemini 4 Argon lands everywhere, while Anthropic builds its own Git to survive Claude's output | daily.dev og:description: Google's Gemini 4 Argon went globally available today, landing in Copilot, Antigravity, and a half-dozen IDEs within hours of launch. Anthropic revealed it had... og:url: https://daily.dev/posts/gemini-4-argon-lands-everywhere-while-anthropic-builds-its-own-git-to-survive-claude-s-output-toun9wwtd og:image: https://api.daily.dev/og/posts/tOUn9wwTD.png og:image:alt: Gemini 4 Argon lands everywhere, while Anthropic builds its own Git to survive Claude's output og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Gemini 4 Argon lands everywhere, while Anthropic builds its own Git to survive Claude's output **[Agentic Digest](https://daily.dev/sources/agents_digest)** · 4 min read · 0 upvotes · 0 comments ## Summary Google's Gemini 4 Argon went globally available today, landing in Copilot, Antigravity, and a half-dozen IDEs within hours of launch. Anthropic revealed it had to build an entirely new version control system, Scion, because Claude-generated code changes grew twelvefold since May and broke its existing Git infrastructure. A Chrome bounty hunter spent $40,000 in GPT-5.6 compute to win a $250,000 Google payout, and LiteLLM patched two gateway flaws serious enough to hand over provider API keys. Meanwhile the 'push to main, skip code review' debate picked up real backers, and a new Carnegie Mellon/Microsoft benchmark shows even the best AI reviewer still misses almost 30% of real bugs. ## Content **TLDR:** Google's Gemini 4 Argon went globally available today, landing in Copilot, Antigravity, and a half-dozen IDEs within hours of launch. Anthropic revealed it had to build an entirely new version control system, Scion, because Claude-generated code changes grew twelvefold since May and broke its existing Git infrastructure. A Chrome bounty hunter spent $40,000 in GPT-5.6 compute to win a $250,000 Google payout, and LiteLLM patched two gateway flaws serious enough to hand over provider API keys. Meanwhile the 'push to main, skip code review' debate picked up real backers, and a new Carnegie Mellon/Microsoft benchmark shows even the best AI reviewer still misses almost 30% of real bugs. --- ## Gemini 4 Argon ships globally with unified 1M token context Google's new flagship model is now generally available after a two-week cyber-defender preview, bringing a unified 1 million token budget shared between input and output. Introductory pricing runs $2/$10 per million input/output tokens through November 30, rising to $4/$20 to match Claude Opus 5.5. It's already live in GitHub Copilot, Antigravity, Android Studio, and Search's AI Mode, and early users are posting results from a 6-million-line Chrome refactor to a detailed 3D pagoda built through multi-agent visual critique loops. The breadth of same-day integration is the real story here, not any single benchmark. [Read more](https://daily.dev/feed-by-ids?id=UqDxppv6j&id=nUUsUhM7A&id=YhVBQnC13&id=wtZAHs61E&id=ouvzwHaDZ&id=Ts9NU3Y3b&id=uyYHsOzaP&id=JHQl5YwUl) ## Anthropic built its own source control because Claude writes too much code Daily merged code changes at Anthropic went from about 2,500 in May to over 30,000 by September, and the company's existing infrastructure couldn't keep up. The fix is Scion, a homemade centralized monorepo system echoing Google's Piper and Meta's Sapling, built alongside a new code review tool, code search, and an IAM system. Claude itself helped write the Git-to-Scion ingestion pipeline, which ended up 20x more memory-efficient and 4x faster than the original. Anthropic says it won't open-source Scion, but expects every AI-heavy engineering org to hit this same wall soon. [Read more](https://daily.dev/feed-by-ids?id=0uLq5tegm&id=9aU5Y6k5F) ## A $40,000 agent run bought a $250,000 Chrome exploit Xbow pointed OpenAI's gpt-5.6-sol at Chrome's source code and, over six days and roughly $40,000 of compute, produced a full-chain exploit crossing two sandbox boundaries through three separate bugs. Google confirmed all three, paid out $250,000 — only the second time it's ever paid that amount for a full-chain bounty — and shipped fixes. Humans still picked the targets and joined the pieces together, so this isn't a model hacking Chrome solo, but the economics are the headline: a budget version of the same idea (an open-source agent called Shannon) found eight real vulnerabilities in a personal Kubernetes app for about $2 in three hours. [Read more](https://daily.dev/posts/bVFGE3w7i) ## LiteLLM flaws let low-privileged users steal provider API keys Check Point disclosed two high-severity bugs in LiteLLM, the AI gateway pulled roughly 97 million times a month: a missing authorization check on /config/update that let any logged-in user read arbitrary files, hijack the admin account, or get remote code execution, and an SSRF flaw that let attackers redirect provider API keys to their own servers. Both are fixed in version 1.83.9. If you're running LiteLLM anywhere near production, this is a patch-now situation, not a read-later one. [Read more](https://daily.dev/feed-by-ids?id=Z1EUe9Ebf&id=Cg5I3UzUB) --- ## Also notable - **CRBugBench shows the best AI reviewer still misses 28% of bugs:** Claude-opus-5 caught 71.9% of human-confirmed bugs in its top 5 comments across 1,800 real pull requests, up from 47.7% for opus-4.6 just five months earlier, but accuracy dropped as much as 25.5 points on one company's private repos. [Read more](https://daily.dev/posts/6Bu014pCf) - **Amp pushes straight to main, no PRs, no review:** Amp's Thorsten Ball says agents now read production logs and watch deploys instead of going through pull request review, a stance reshared without caveats by several well-known dev tool builders. [Read more](https://daily.dev/posts/0aAMGzaWk) - **Giving multimodal agents tools makes them worse at refusing harmful requests:** New NVIDIA research accepted at NeurIPS 2026 found refusal failures rise by up to 68.7% relative (17.7% average) once a model gets tools, across every model tested including Claude Opus 4.6/4.7 and Gemini Agentic Vision. [Read more](https://daily.dev/posts/4NjYTUqS9) - **Character.AI files Chapter 11 as Disney bids $20M for its IP:** Character.AI listed $50-100M in assets and liabilities in its Delaware bankruptcy filing, with Disney as lead bidder for a non-exclusive license to its IP, five years after the company's $2.7B Google licensing deal. [Read more](https://daily.dev/posts/tK5M30b5g) - **Codex CLI silently runs a persistent background daemon:** Since version 0.157.0, opening an interactive Codex session starts a self-updating ~374MB daemon that keeps running after you close the terminal, with analytics on by default and no per-client isolation. [Read more](https://daily.dev/posts/uC5UkARla) ## Similar posts on daily.dev - [Gemini 4 Argon in GitHub Copilot](https://daily.dev/posts/gemini-4-argon-in-github-copilot-nuusuhm7a) · GitHub Changelog · 0 upvotes · 0 comments - [Gemini 4 Argon debuts with a unified 1M input/output token limit](https://daily.dev/posts/gemini-4-argon-debuts-with-a-unified-1m-input-output-token-limit-uqdxppv6j) · The New Stack · 5 upvotes · 2 comments - [Gemini 4 Argon: Google’s New AI Is Built to Think and Work for Much Longer](https://daily.dev/posts/gemini-4-argon-google-s-new-ai-is-built-to-think-and-work-for-much-longer-ibbgfhpdj) · Medium · 0 upvotes · 0 comments --- Tags: [#claude](https://daily.dev/tags/claude), [#anthropic](https://daily.dev/tags/anthropic), [#google-gemini](https://daily.dev/tags/google-gemini), [#version-control](https://daily.dev/tags/version-control), [#appsec](https://daily.dev/tags/appsec) [View this post on daily.dev](https://daily.dev/posts/gemini-4-argon-lands-everywhere-while-anthropic-builds-its-own-git-to-survive-claude-s-output-toun9wwtd) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"DiscussionForumPosting","mainEntityOfPage":"https://daily.dev/posts/gemini-4-argon-lands-everywhere-while-anthropic-builds-its-own-git-to-survive-claude-s-output-toun9wwtd","headline":"Gemini 4 Argon lands everywhere, while Anthropic builds its own Git to survive Claude's output","text":"Google's Gemini 4 Argon went globally available today, landing in Copilot, Antigravity, and a half-dozen IDEs within hours of launch. Anthropic revealed it had to build an entirely new version control system, Scion, because Claude-generated code changes grew twelvefold since May and broke its existing Git infrastructure. A Chrome bounty hunter spent $40,000 in GPT-5.6 compute to win a $250,000 Google payout, and LiteLLM patched two gateway flaws serious enough to hand over provider API keys. Meanwhile the 'push to main, skip code review' debate picked up real backers, and a new Carnegie Mellon/Microsoft benchmark shows even the best AI reviewer still misses almost 30% of real bugs.","url":"https://daily.dev/posts/gemini-4-argon-lands-everywhere-while-anthropic-builds-its-own-git-to-survive-claude-s-output-toun9wwtd","datePublished":"2026-10-14T04:17:50.507Z","dateModified":"2026-10-14T04:18:10.621Z","author":{"@type":"Organization","name":"Agentic Digest","logo":"https://media.daily.dev/image/upload/s--V91DY4ls--/f_auto,q_auto/v1772617267/logos/agents_digest","url":"https://daily.dev/sources/agents_digest"},"interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"isPartOf":{"@type":"WebPage","url":"https://daily.dev/sources/agents_digest","name":"Agentic Digest"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Agentic Digest","item":"https://daily.dev/sources/agents_digest"},{"@type":"ListItem","position":3,"name":"Gemini 4 Argon lands everywhere, while Anthropic builds its own Git to survive Claude's output"}]} ```