{"slug": "over-1500-in-tests-found-rtk-output-cuts-may-not-lower-ai-coding-costs", "title": "Over $1,500 in tests found RTK output cuts may not lower AI coding costs", "summary": "Benchmarks costing over $1,500 found that RTK, a terminal output compression tool, cut costs by only 5% for Claude Code with Fable 5.0 while raising costs by 5% for OpenCode with DeepSeek V4 Pro 0813, according to a Quesma blog post. The results indicate that reducing token count does not consistently translate into lower AI coding expenses, and the post advises engineers running LLMs in production to test RTK in their own workflows before relying on it.", "body_md": "[Hacker News](https://quesma.com/blog/does-rtk-make-ai-coding-cheaper/)\n\n### Over $1,500 in tests found RTK output cuts may not lower AI coding costs\n\nWhich summary reads better? Pick one — models revealed after.Both summaries are AI-generated.\n\nRTK, a popular tool for compressing terminal output, resulted in only a 5% cost reduction for Claude Code with Fable 5.0 and actually increased costs by 5% for OpenCode with DeepSeek V4 Pro 0813 in recent benchmarks. This discrepancy highlights that simply reducing token count does not directly translate to cost savings, potentially impacting the cost-effectiveness of AI coding workflows. Engineers running LLMs in production should reassess the value of such token-saving tools in their specific use cases.\n\nRTK's terminal output compression showed mixed results: costs dropped by 5% with Claude Code but increased by 5% with OpenCode, highlighting that token savings don’t consistently translate to lower expenses. Engineers must rigorously test RTK in their specific workflows, as its impact varies by model and task, potentially increasing costs or reducing efficiency rather than delivering promised savings.", "url": "https://wpnews.pro/news/over-1500-in-tests-found-rtk-output-cuts-may-not-lower-ai-coding-costs", "canonical_source": "https://www.snipvote.com/story/cmtzhp7k2000afsawj6z4hj0u", "published_at": "2026-09-13 07:57:33.699146+00:00", "updated_at": "2026-09-13 07:57:35.190857+00:00", "lang": "en", "topics": ["ai-tools", "developer-tools", "large-language-models", "ai-products"], "entities": ["RTK", "Claude Code", "Fable 5.0", "OpenCode", "DeepSeek V4 Pro 0813", "Quesma", "Hacker News"], "alternates": {"html": "https://wpnews.pro/news/over-1500-in-tests-found-rtk-output-cuts-may-not-lower-ai-coding-costs", "markdown": "https://wpnews.pro/news/over-1500-in-tests-found-rtk-output-cuts-may-not-lower-ai-coding-costs.md", "text": "https://wpnews.pro/news/over-1500-in-tests-found-rtk-output-cuts-may-not-lower-ai-coding-costs.txt", "jsonld": "https://wpnews.pro/news/over-1500-in-tests-found-rtk-output-cuts-may-not-lower-ai-coding-costs.jsonld"}}