![\"\"](\"https:\/\/trackstreamai.com\/wp-content\/uploads\/2026\/06\/\u622a\u5c4f2026-06-18-10.31.13.png\")
<\/figure>\n\n\n\nRecently, institutions like the University of Michigan and Stanford have published a paper. They tested 8 of the strongest AIs on the market, assigning them the same task and calculating the costs in detail. The AI that consumed the most resources performed the worst, and most of the cost was on the input side, with an input Token to output Token ratio of 107:1.<\/p>\n\n\n\n
This paper is full of experimental data and cost analysis, which is hard for ordinary people to understand. So I created an AI agent to help me break it down. If you\u2019re new to building and running local AI agents and want to avoid runaway token spending, OPENCLAW BOOK FOR BEGINNERS<\/strong><\/em> <\/a>offers a straightforward hands-on foundation to set up efficient open-source agent workflows on your own machine. You can grab the Kindle edition directly on Amazon<\/a><\/strong>.<\/p>\n\n\n\n When you chat with AI, ask a question, or let it write a copy, it only takes hundreds to thousands of Tokens, which costs just a few cents. But the paper also found that for AI agents that work autonomously, completing a task consumes 1200 times more Tokens than a regular code Q&A and 3500 times more than a single-round reasoning task.<\/p>\n\n\n\n What does 1200 times mean? It\u2019s like asking your assistant to buy a bottle of water, but they bring back the entire supermarket. You think they\u2019re helping you work, but most of the time, they\u2019re flipping through things, re-reading the same page, and checking information they already know. With an input-output ratio of 107:1, most of their time isn\u2019t spent working, but re-reading your requirements, materials, previous conversations, and error messages. OpenClaw\u2019s local-first design mitigates this waste by letting you fully control context window and tool invocation rules from the start, as covered simply in the beginner\u2019s guide.<\/p>\n\n\n\n It\u2019s like a very verbose employee. If you ask them to write a proposal, they\u2019ll first spend half a day going through the company\u2019s historical archives and then your chat records from the past three years. Finally, they spend ten minutes writing the proposal, and all the Token costs are on you. What\u2019s more frustrating is that \u201cspending more\u201d doesn\u2019t mean \u201cdoing better\u201d. The paper found that some models burned an average of 1.5 million more Tokens than the most cost-effective GPT-5, but the effect wasn\u2019t better. Accuracy hits a ceiling at a certain point, and any additional Tokens beyond that are wasted.<\/p>\n\n\n\n AI agents aren\u2019t better just because they work harder; working too hard is just a waste. It\u2019s like having a meeting where you make a decision in the first 20 minutes, but then spend the next two hours going in circles. The most darkly humorous part is: does AI know how much it\u2019s spending? The paper tried to let AI estimate the cost of completing a task before executing it, but all 8 models failed. The highest correlation between prediction and actual cost was only 0.39 (with 1 being a perfect match), so 0.39 is basically a blind guess. All models underestimated the cost; none overestimated. They thought it would cost 10 yuan, but actually spent 50 yuan. AI also severely underestimates its own reading costs; it has no idea how much it\u2019s flipping through. It\u2019s like someone who doesn\u2019t keep accounts\u2014they\u2019re shocked when they see the bill at the end of the month.<\/p>\n\n\n\n What\u2019s \u201cdifficult\u201d for humans and what\u2019s \u201cexpensive\u201d for AI are two different things. A task you think is simple might have AI burning money like crazy, while a task you think is extremely difficult might be done by AI in no time. This paper, in the end, exposes a blind spot in the industry: the cost logic changes completely when AI moves from chatting to actually working. Now, everyone is competing to see who has the strongest model and the most parameters, but no one cares about how much wasted money is being spent.<\/p>\n\n\n\n The future competition for AI agents won\u2019t be about who has the biggest model, but who is better at saving money. Speaking of which, we\u2019ve genetically modified our \u201cLobster\u201d AI. It has reasonable computing power scheduling\u2014using lightweight models for simple tasks and stronger models for complex tasks, balancing effect and cost. You don\u2019t have to choose by yourself; the system schedules for you, saving 50%-100% of Token consumption. This core logic of tiered model routing and context pruning is also covered in plain language within OpenClaw\u2019s beginner guide for anyone building personal multi-agent systems.<\/p>\n\n\n\n In the end, in the AI agent track, the ability to save costs is the real skill. Which is the most expensive AI you\u2019ve used? Let\u2019s chat in the comments. If you want to start building cost-efficient local AI agents without dense technical jargon, head to Amazon<\/strong> to pick up the OpenClaw beginner book here.<\/p>\n\n\n\n Affiliate Disclosure: Some links in this post are affiliate links. I may earn a small commission if you purchase through them, at no extra cost to you.<\/em> When AI Does the Same Task Twice, Token Costs Can Differ by 30x If you let AI do the same task twice, the money spent can differ by 30 times. What\u2019s worse, the one that costs 30 times more might even do a worse job than the cheaper one. Recently, institutions like the University of … Read more<\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[76,61,56,77],"class_list":["post-55","post","type-post","status-publish","format-standard","hentry","category-uncategorized","tag-ai-cost-efficiency","tag-ai-model-cost-comparison","tag-ai-research-token-usage","tag-token-optimization-for-ai"],"_links":{"self":[{"href":"https:\/\/trackstreamai.com\/index.php?rest_route=\/wp\/v2\/posts\/55","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/trackstreamai.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/trackstreamai.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/trackstreamai.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/trackstreamai.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=55"}],"version-history":[{"count":1,"href":"https:\/\/trackstreamai.com\/index.php?rest_route=\/wp\/v2\/posts\/55\/revisions"}],"predecessor-version":[{"id":57,"href":"https:\/\/trackstreamai.com\/index.php?rest_route=\/wp\/v2\/posts\/55\/revisions\/57"}],"wp:attachment":[{"href":"https:\/\/trackstreamai.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=55"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/trackstreamai.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=55"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/trackstreamai.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=55"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
<\/p>\n","protected":false},"excerpt":{"rendered":"