---
title: 'In the News: October 13, 2026 (Midday)'
description: 'Google details an 11-hour unattended Gemini 4 Argon run on its ADK repo: repo rulefiles, a holdout benchmark, and four of six changes rejected in review.'
canonical_url: 'https://darkfactory.dev/news/2026-10-13-midday'
markdown_url: 'https://darkfactory.dev/news/2026-10-13-midday.md'
collection: news
date_published: '2026-10-13T13:05:00-04:00'
date_modified: '2026-10-13T13:05:00-04:00'
---

# In the News: October 13, 2026 (Midday)


Google released Gemini 4 Argon today and published an engineering write-up of an overnight run on its own agent framework. The write-up is the more useful of the two for anyone running agents against an existing codebase.

## 1. Google ran Gemini 4 Argon for 11 hours on its ADK repo and rejected four of its six changes

**[Optimizing a 200,000-line Python codebase overnight with Gemini 4 Argon](https://developers.googleblog.com/optimizing-a-200000-line-python-codebase-overnight-with-gemini-4-argon/)** · George Weale, Bo Yang and Andrew Biernat, Google Cloud · Google Developers Blog, October 13, 2026

The target was google/adk-python, roughly 205,000 lines across nearly 800 source files with nearly 17,000 unit tests. A first two-hour run, given one paragraph of goal and no repository instructions, found real bottlenecks but altered the public JSON schema of ADK's authentication model, missed mock edge cases, and ran only 216 tests. For the second run the team put its maintainer guidelines in the repository's .agents/skills directory and used Antigravity's /goal mode, which keeps the goal across context compactions. That run went 11 hours, read 12 guideline files first, ran all 16,892 tests, and opened a pull request with six optimizations across 32 files. Reviewers shipped two and rejected four, including a change that sped up a synthetic benchmark by 17% but relied on a weak memory reference that could return stale data across sessions.

The merged change touches 22 files and adds 17 test functions. On a holdout benchmark the model never saw, Google reports about 40% less in-memory session preparation overhead in long sessions and 21% faster cold imports. Time from a new process to the first model request stayed flat, at 3.07 s against 3.09 s. The authors also say 25 internal Google test targets failed on an early lazy-import attempt even though every open-source test passed. Their summary: "the agent explores, the maintainer subtracts."

**Why it matters:** The transferable part is the setup, not the speedup. The same model went from a run that changed a public schema to one that followed repo conventions once the rules were written as files and the goal survived compaction. The holdout benchmark and human review caught what the agent's own tests did not. If you run agents overnight, write your review conventions into the repo, keep a benchmark the agent cannot see, and budget reviewer time for cutting changes. This is one run on Google's own framework with Google's own model and tooling, and the write-up cites no independent replication.

## 2. Google releases Gemini 4 Argon at $2 and $10 per million tokens, with a 1M-token limit

**[Gemini 4 Argon: Built to solve your hardest problems](https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-argon/)** · Koray Kavukcuoglu, Google DeepMind · Google blog, October 13, 2026

Argon is available through the Gemini API under the model ID gemini-4-argon, and in Google Antigravity for AI Ultra subscribers. Google lists an introductory price of $2 per million input tokens and $10 per million output tokens, with cached input 95% off, and a 1M unified input-output token limit with a new continuation protocol that it documents in the developer guide. The post reports that Google teams using Argon in Antigravity increased agentic code submissions by more than 35% and cut code rollbacks by nearly a third. It also says shortcut attempts on coding benchmarks fell to 2%, the lowest of any Gemini model, and that attack success on the Gray Swan indirect prompt injection benchmark fell to 0.7% across 15 adaptive attempts. All of these are Google's own measurements. Beyond a #1 ranking on the Vals Index, the post gives no coding benchmark scores.

**Why it matters:** The price and token limit are the facts you can act on this month. Rerun your own agent evals against gemini-4-argon before relying on the shortcut and rollback figures, since they come from Google's internal benchmarks and teams. Read the continuation protocol in the developer guide before moving a harness that assumes an ordinary output limit.