Rohan Paul· @rohanpaul_ai · X·· 2026-07-13AI 评分42
AI 导读
Grok 4.5 在浏览器智能体（browser-use agents）评测中进入顶级梯队，得分超过 GPT-5.6-Sol，接近 Claude Opus。由于缓存输入成本较高，其总体使用成本仅比 Opus 便宜约 10%，但速度略快。Rohan Paul 指出，前沿模型质量正在快速压缩，目前已有多个模型达到 Opus 级别性能。
正文 · 原文
Grok 4.5 has moved into the top tier for browser-use agents, scoring above GPT-5.6-Sol and landing close to Claude Opus.
frontier model quality is compressing fast. We have now quite a few option for roughly Opus-class performance.
I take it all back. We just got access to eval Grok 4.5 and it has landed above GPT-5.6-Sol and just shy of Opus for browser use. Because cache input is expensive, the overall cost is only 10% cheaper than opus. Its overall a bit faster. We have another opus-class model in the competition在 X 查看被引用的帖子
来源：Rohan Paul · x.com