TechnologyPhoto: NDTV TechGemini 4 Carbon Trails Claude Opus 5.5 in Agentic Tasks but Is More Cost-Efficient
Google's Gemini 4 Carbon AI model ranked second in agentic task performance on SemiAnalysis' GDPx benchmark, scoring 10.9 percentage points compared to Anthropic's Claude Opus 5.5 at 12 points. Despite trailing slightly, Gemini 4 Carbon was much more cost-efficient, completing tasks at $16.29 each with 594,000 output tokens, while Claude Opus 5.5 cost $51.41 per task with 548,000 tokens.
NDTV Tech·


