RTK reports token savings, but our cost benchmarks disagree
Hacker News 热议:RTK reports token savings, but our cost benchmarks disagree(134 赞 / 67 评论,来源 quesma.com)
一句话概要
We tested RTK (Rust Token Killer) with Claude Code on Fable 5.0, and OpenCode with DeepSeek V4 Pro 0813 on Terminal-Bench 2.1.
原文开头节选
Download PNG Hacker News front page · 11 September 2026 RTK (Rust Token Killer) filters and compresses terminal output before the AI agent reads it. With over 79k GitHub stars today, RTK is one of the most popular tools to make AI coding cheaper.
One X post saying RTK could cut Claude Code tokens by up to 60% reached 313K views.
RTK cuts up to 90% of the bash output your agent reads. […] it is not the same as cutting your bill by 90%.
So “less terminal output” is not the same as “cheaper AI coding”. It can help, be a no-op, or backfire (more turns or lower quality). In this post, we present our findings after several days and over $1,500 spent on tokens.
(以上为原文节选,完整内容见下方”原文来源”)
这条动态今日登上 Hacker News 首页(134 赞 / 67 评论,来源 quesma.com)。技术雷达每日自动聚合 AI 工程、后端架构、DevOps 方向的前沿动态;相关工程落地可浏览下方的相关服务与延伸阅读,或直接与我们团队交流。
原文来源: Hacker News