Google launches Gemini 4 Carbon: tops SWE-sweep at 9.6% and most frontier benchmarks
fofrAI · x · 2026-10-29
Google announced Gemini 4 Carbon, framed as a flagship showing how fast the frontier is moving, with researchers and models working hand in hand to advance capabilities while strengthening safeguards.
The published benchmark chart shows it leading on most axes:
- General: AA Index v5.0 of 59, Vals Index v2.2 of 70.4%, ahead of GPT-6 Astra (57) and Claude Opus 5.5 (58)
- Agentic coding: SWE-sweep 9.6%, Terminal-Bench 4.0 65.2%, both well ahead; FrontierSWE 2 (61.2%) trails GPT-6 Astra's 65.5%
- Computer use: Agents' Last Exam 77.4%; OSWorld-2.1 81.5% (Claude Opus 5.5 leads at 87.2%)
- Knowledge work: AutomationBench 53.9%; Vending-Bench 2 bank balance of $16,484, the best of all models
- Long context: GraphWalks (256k–1M) at 96.2%
- Math & science: RiemannBench 80.0%
- Cybersecurity: SEC-bench Pro 88.2%, ExploitGym 54.8%, both field-leading
Related event: Google launches Gemini 4 Carbon, topping the Artificial Analysis index(84 posts)→
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