--- title: OpenAI Just Made Analytics 10x Cheaper | daily.dev description: OpenAI's 80% price cut for GPT 5.6 Luna has made AI-powered analytics dramatically cheaper, enabling accurate SQL generation for under half a penny per answer.... canonical: https://daily.dev/posts/openai-just-made-analytics-10x-cheaper-zo6zoqpvf twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: OpenAI Just Made Analytics 10x Cheaper | daily.dev og:description: OpenAI's 80% price cut for GPT 5.6 Luna has made AI-powered analytics dramatically cheaper, enabling accurate SQL generation for under half a penny per answer.... og:url: https://daily.dev/posts/openai-just-made-analytics-10x-cheaper-zo6zoqpvf og:image: https://api.daily.dev/og/posts/zo6ZoQpvf.png og:image:alt: OpenAI Just Made Analytics 10x Cheaper og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # OpenAI Just Made Analytics 10x Cheaper **[MotherDuck](https://daily.dev/sources/motherduck)** · 8 min read · 2 upvotes · 0 comments ## Summary OpenAI's 80% price cut for GPT 5.6 Luna has made AI-powered analytics dramatically cheaper, enabling accurate SQL generation for under half a penny per answer. Luna on max effort now matches or beats prior top models on agentic SQL benchmarks at a fraction of the cost. This shift changes the calculus for data teams: the bottleneck moves from LLM cost to database latency, making fast analytical engines like DuckDB/MotherDuck more valuable. Key recommendations include switching from large models on low effort to Luna on max, investing in richer business context layers, running evals far more frequently, and designing eval workflows to be model-agnostic to avoid vendor lock-in. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How much did OpenAI cut the price of GPT 5.6 Luna and what was the effect on SQL benchmark performance? OpenAI cut GPT 5.6 Luna's price by 80%. On an agentic SQL benchmark, Luna running at max effort then outperformed Gemini-3-Flash on low effort settings, holding 99.8% accuracy at a 5x lower price point than the prior champion, and semantic modeling benchmark costs dropped 10x due to the price cut plus model caching improvements. _Developers tuning model choice for SQL-generation workloads can follow analytics and AI pricing shifts like this on daily.dev._ ### Why does a faster AI model shift the performance bottleneck to the database in agentic analytics workflows? When an LLM call took tens of seconds per turn, database speed barely mattered because inference dominated total latency. With models like GPT 5.6 Luna responding much faster, the analytical database becomes the new bottleneck, since analytical engines can be up to 1000x faster than transactional databases on analytically shaped queries, making low-latency serverless analytical stores critical for responsive agent experiences. _Teams designing low-latency agentic data products can track this shift from AI latency to database latency on daily.dev._ ### Why do AI agent data teams need more than traditional data quality checks for evals? Traditional data quality checks only verify computable invariants like no duplicate customer IDs or null order prices, but agentic analysis requires evaluating whether an agent can answer natural language questions correctly using business context and a live database connection. Cheaper inference (evals got roughly 5x cheaper) makes running these richer evals continuously feasible, catching context regressions and business logic drift, such as new discount programs breaking revenue calculations. _Engineers building eval frameworks for AI-driven analytics can find related practices on daily.dev._ --- Tags: [#openai](https://daily.dev/tags/openai), [#agentic-ai](https://daily.dev/tags/agentic-ai), [#duckdb](https://daily.dev/tags/duckdb) [View this post on daily.dev](https://daily.dev/posts/openai-just-made-analytics-10x-cheaper-zo6zoqpvf) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"OpenAI Just Made Analytics 10x Cheaper","url":"https://daily.dev/posts/openai-just-made-analytics-10x-cheaper-zo6zoqpvf","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/openai-just-made-analytics-10x-cheaper-zo6zoqpvf"},"datePublished":"2026-08-03T11:51:41.457Z","dateModified":"2026-09-14T06:52:32.559Z","description":"OpenAI's 80% price cut for GPT 5.6 Luna has made AI-powered analytics dramatically cheaper, enabling accurate SQL generation for under half a penny per answer....","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/77343e54926b199d9b9237cc815c1c3a?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/77343e54926b199d9b9237cc815c1c3a?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"MotherDuck","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"MotherDuck","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/f084d6835eef4c2b83f889f9a4e85ed6","url":"https://daily.dev/sources/motherduck"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/openai-just-made-analytics-10x-cheaper-zo6zoqpvf","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":2},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"openai,agentic-ai,duckdb","timeRequired":"PT8M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"MotherDuck","item":"https://daily.dev/sources/motherduck"},{"@type":"ListItem","position":3,"name":"OpenAI Just Made Analytics 10x Cheaper"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/openai-just-made-analytics-10x-cheaper-zo6zoqpvf#faq","mainEntity":[{"@type":"Question","name":"How much did OpenAI cut the price of GPT 5.6 Luna and what was the effect on SQL benchmark performance?","acceptedAnswer":{"@type":"Answer","text":"OpenAI cut GPT 5.6 Luna's price by 80%. On an agentic SQL benchmark, Luna running at max effort then outperformed Gemini-3-Flash on low effort settings, holding 99.8% accuracy at a 5x lower price point than the prior champion, and semantic modeling benchmark costs dropped 10x due to the price cut plus model caching improvements. Developers tuning model choice for SQL-generation workloads can follow analytics and AI pricing shifts like this on daily.dev."}},{"@type":"Question","name":"Why does a faster AI model shift the performance bottleneck to the database in agentic analytics workflows?","acceptedAnswer":{"@type":"Answer","text":"When an LLM call took tens of seconds per turn, database speed barely mattered because inference dominated total latency. With models like GPT 5.6 Luna responding much faster, the analytical database becomes the new bottleneck, since analytical engines can be up to 1000x faster than transactional databases on analytically shaped queries, making low-latency serverless analytical stores critical for responsive agent experiences. Teams designing low-latency agentic data products can track this shift from AI latency to database latency on daily.dev."}},{"@type":"Question","name":"Why do AI agent data teams need more than traditional data quality checks for evals?","acceptedAnswer":{"@type":"Answer","text":"Traditional data quality checks only verify computable invariants like no duplicate customer IDs or null order prices, but agentic analysis requires evaluating whether an agent can answer natural language questions correctly using business context and a live database connection. Cheaper inference (evals got roughly 5x cheaper) makes running these richer evals continuously feasible, catching context regressions and business logic drift, such as new discount programs breaking revenue calculations. Engineers building eval frameworks for AI-driven analytics can find related practices on daily.dev."}}]} ```