Gemini 4 Carbon (High) hits #2 in Agent Arena at $1.80/task, 72% cheaper than Claude Opus 5.5
arena · x · 2026-10-29
LMArena announced that Gemini 4 Carbon (High) from Google DeepMind landed on the Pareto frontier across Agent Arena, Code Arena: WebDev, and Text Arena.
- Ranks #2 in Agent Arena with +16.2% net improvement at a median cost of $1.80 per task
- That's 72% cheaper than the #1 model, Claude Opus 5.5 (Max)
The result positions the new model as a strong cost-performance option for agentic workloads, nearly matching top-tier performance at a fraction of the price.
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