ModelsMulti-source
SemiAnalysis 发布智能体基准 GDPx，最强模型仅得 12%
Why we picked it
SemiAnalysis 弄了个智能体新基准 GDPx，最强模型才拿 12% 分，Pichai 还点名 Carbon 的成本表现，想看模型真实水平差距的可以瞅瞅。
SemiAnalysis 推出 GDPx 基准，用于衡量模型在高价值领域中的智能体能力。该基准难度达到前沿水平，目前表现最好的模型只拿到 12%（7 题中答对 1 题）。Sundar Pichai 转发时特别提到 Carbon 模型的成本线，暗示其在成本与性能之间存在较大提升空间。
Original · Sundar Pichai
Great to see high quality benchmarks with headroom and look at that cost line for Carbon🚀 SemiAnalysis @SemiAnalysis_ Today we’re releasing GDPx, a benchmark for measuring agentic capabilities across the most valuable domains. GDPx is a true frontier-level benchmark: even the best model scores just 12% (1/7)🧵 🔗 View Quoted Tweet 💬 348 🔄 267 ❤️ 4285 👀 487906 📊 382 ⚡
