In @googlechrome, Gemini 4 Argon helped our team separate core webpage handling from a 6-million-line legacy folder of browser features in Chromium’s 40-million-line codebase — completing months of work in just weeks.
The result for Chrome's billions of users: eliminating an ser crashes, and letting live webpages move fluidly between tabs, side panels, and floating windows without reloading.
The result for Chrome's billions of users: eliminating an ser crashes, and letting live webpages move fluidly between tabs, side panels, and floating windows without reloading.
13
10
115
14,363
Software engineers at Google using Argon in @antigravity are shipping high-quality code faster than ever.
Over the past few weeks, agentic code submissions increased by more than 35% — while cutting code rollbacks by nearly a third.
Read more about the impact of Gemini 4 Argon goo.gle/4hPJbGo
Over the past few weeks, agentic code submissions increased by more than 35% — while cutting code rollbacks by nearly a third.
Read more about the impact of Gemini 4 Argon goo.gle/4hPJbGo
3
3
63
19,045
For Google's open-source Agent Development Kit (ADK), the team gave Argon a skills rulebook and one goal: eliminate framework overhead without breaking existing apps.
Working autonomously overnight, Argon surfaced optimizations that cut in-memory session prep time by 40% in long ld imports by 21%.
See it live in google/adk-python → goo.gle/4eAtLDA
Working autonomously overnight, Argon surfaced optimizations that cut in-memory session prep time by 40% in long ld imports by 21%.
See it live in google/adk-python → goo.gle/4eAtLDA
Oct 15, 2026 · 2:55 PM UTC
3
0
37
7,513
Our researchers also used Argon in @antigravity to optimize xTB, one of the most-cited open-source quantum chemistry packages for predicting molecules' 3D shapes and screening drug and materials candidates.
They made xTB up to 2.3x faster at predicting molecular structures and r search, contributing the changes to the official xTB repository on GitHub → goo.gle/4s6dko0
They made xTB up to 2.3x faster at predicting molecular structures and r search, contributing the changes to the official xTB repository on GitHub → goo.gle/4s6dko0
1
1
30
7,227
Gemini 4 Argon is also speeding up our @googlegemma open models — making them nearly three times as fast.
In just a few days, Argon optimized Gemma 4 E2B to achieve a 2.85x inference speedup on an A100 via vLLM, jumping from ~37,900 to over 108,000 tokens per second with 100% token parity and zero quality drop.
Here’s that speedup in action, generating 64 live Python simulations at once ↓
x.com/osanseviero/st…
In just a few days, Argon optimized Gemma 4 E2B to achieve a 2.85x inference speedup on an A100 via vLLM, jumping from ~37,900 to over 108,000 tokens per second with 100% token parity and zero quality drop.
Here’s that speedup in action, generating 64 live Python simulations at once ↓
x.com/osanseviero/st…
3
3
80
24,514
@googlegemmaing legacy Chromium code to speeding up our @GoogleGemma @googlegemman is built to solve highly complex coding and agentic problems at Google and beyond.
We’re also opening applications for researchers and cyber defenders:
🔹 Scientists & researchers: Apply for a chance to get @Antigravity’s highest usage tier free for 6 months (subject to eligibility & selection) → goo.gle/4z0OCrZ
🔹 Cybersecurity defenders: Apply to join the Fairwind Program → goo.gle/3TUNMOj
Learn more about Gemini 4 Argon ↓ goo.gle/4jfSznG
We’re also opening applications for researchers and cyber defenders:
🔹 Scientists & researchers: Apply for a chance to get @Antigravity’s highest usage tier free for 6 months (subject to eligibility & selection) → goo.gle/4z0OCrZ
🔹 Cybersecurity defenders: Apply to join the Fairwind Program → goo.gle/3TUNMOj
Learn more about Gemini 4 Argon ↓ goo.gle/4jfSznG
3
1
48
22,343
@Google I followed the commit link. The full 158-turn run takes 14% less time with dummy models/tools; the ~40% is the long-session prep saving. Those are two different waits, and it’s useful to see both timed.
0
0
0
7