Gemini 4 Carbon Frontier AI Model Falls Slightly Behind Claude Opus 5.5 in Agentic Capabilities, but Not in Cost Efficiency

Google’s Gemini 4 Carbon was the second best at performing agentic tasks on SemiAnalysis' recently released GDPx benchmark test. The frontier AI model managed to score 10.9 percentage points, which was slightly lower than the 12-percentage-point score of Anthropic’s Claude Opus 5.5. While Google’s AI model came in second position, it was significantly more cost-...