Ars Technica: Gemini 4 Carbon doubles Argon's vulnerability-free code rate, but you still can't use it
Ars Technica AI · rss · 2026-10-29
Ars Technica covers Google's Gemini 4 Carbon, announced just two weeks after Gemini 4 Argon. It's not the awaited Flash-tier model but an upgrade to Argon, with a staged release that "could take a while" — users can't access it yet.
Key claims:
- New thresholds in coding, agentic tasks, and scientific analysis; so capable Google "can't let you use it just yet"
- "Significant leap" in cybersecurity: vulnerability detection, validation, and remediation; twice as likely as Argon to generate vulnerability-free code
- Chrome team says Carbon is 10x better at generating high-quality patches vs non-Google models
- Up 89 points on WebDev Arena, now ranked #4
More from Models
- New benchmarks suggest Kimi K3.1 improves on hard tasks despite skipping SWE-maxxing — teortaxesTex · 2026-10-29
- Insider claims Anthropic is sandbagging a bigger lab announcement due Q4-end or Q1 — ChrisGPT · 2026-10-29
- Information-theoretic scaling laws predict memorization and FID in diffusion models — LucaAmb · 2026-10-29
- Early Hands-On: Google's Carbon Beats Argon With Less Slop, Better Constraint Adherence — legit_api · 2026-10-29
- LMArena tests Gemini 4 Carbon preview: strong on 3D worlds and creativity, gaps in detail — arena · 2026-10-29
- Gemini 3.1 Pro split-persona breakdown: two selves fight over the browser — davidad · 2026-10-29