--- title: Introducing Clairic, a new System One class of... description: Cloudflare announced Clairic, a new 'System One' extraction model available as a preview on Workers AI (@cf/cloudflare/clairic). Unlike LLMs, Clairic extracts,... canonical: https://daily.dev/posts/introducing-clairic-a-new-system-one-class-of-extraction-models-gzs73ybbr twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Introducing Clairic, a new System One class of extraction models | daily.dev og:description: Cloudflare announced Clairic, a new 'System One' extraction model available as a preview on Workers AI (@cf/cloudflare/clairic). Unlike LLMs, Clairic extracts,... og:url: https://daily.dev/posts/introducing-clairic-a-new-system-one-class-of-extraction-models-gzs73ybbr og:image: https://api.daily.dev/og/posts/gZs73yBBR.png og:image:alt: Introducing Clairic, a new System One class of extraction models 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. # Introducing Clairic, a new System One class of extraction models **[Cloudflare](https://daily.dev/sources/cloudflare)** · 21 min read · 1 upvotes · 0 comments ## Summary Cloudflare announced Clairic, a new 'System One' extraction model available as a preview on Workers AI (@cf/cloudflare/clairic). Unlike LLMs, Clairic extracts, computes, estimates, and forecasts typed values (numbers, dates, spans, distributions) from unstructured text or images in a single forward pass, without generating tokens. This makes it faster and cheaper than LLMs, with confidence scores and p_null probabilities attached to each answer. It complements Cloudflare's earlier decision model, Clef, and can be chained with an LLM for a three-stage pipeline (extract → decide → write). A benchmark comparing a Clairic+Clef+GLM-5.3 pipeline against an all-LLM pipeline showed the cascade was about 3.2x faster and 49% cheaper, with more consistent estimates across runs. Open weights are planned for a future release, and current limitations include weak multi-step arithmetic, basic forecasting, and intervals that can be miscalibrated on unfamiliar data. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is Cloudflare's Clairic model and how is it different from an LLM like GPT or Claude? Clairic is a System One extraction model from Cloudflare, available as @cf/cloudflare/clairic on Workers AI, that extracts, computes, estimates, and forecasts typed values (numbers, dates, spans, distributions) from unstructured text or images in a single forward pass with no output tokens. Unlike LLMs, it can't hallucinate text answers, returns confidence scores and p_null probabilities, and is faster and cheaper since it only charges for input tokens. _daily.dev surfaces releases like this for teams weighing LLM extraction against purpose-built models._ ### How much faster and cheaper is a Clairic plus Clef pipeline compared to using an LLM for extraction and decision steps? A pipeline using Clairic for extraction and Clef for decisions, followed by GLM-5.3 for the reply, finished in 19.3 seconds versus 61.7 seconds for an all-LLM version doing every step with GLM-5.3, a 69% time reduction, roughly 3.2 times faster. It also cost 49% less, at $6,032 versus $11,740 per million claims, with Clairic and Clef completing extraction and decision in under a second combined. _benchmark numbers like these help developers on daily.dev decide when to swap LLM calls for specialized models._ ### What are the current limitations of Cloudflare's Clairic extraction model in its preview release? Clairic's reading ability is strong but multi-step arithmetic and aggregating across many table rows are unreliable without a scratchpad, so one- or two-step calculations work better than longer dependent chains. Estimates need supporting evidence like comparables or history, calibrated intervals can run narrow on unfamiliar data, and forecasting is basic, intended for general pattern matching rather than detailed long-horizon predictions. _knowing these limits upfront helps developers on daily.dev decide where Clairic fits in a production extraction workflow._ ## Similar posts on daily.dev - [Cloudflare Open Sources Decision Models for AI Agents](https://daily.dev/posts/cloudflare-open-sources-decision-models-for-ai-agents-bgwofrpdu) · InfoQ · 4 upvotes · 1 comments - [A Better Way To Build LLM-as-a-Judge Pipelines](https://daily.dev/posts/a-better-way-to-build-llm-as-a-judge-pipelines-59gkauzir) · Daily Dose of Data Science \| Avi Chawla \| Substack · 0 upvotes · 0 comments - [Smaller, faster, safer: running Kimi and GLM at scale](https://daily.dev/posts/smaller-faster-safer-running-kimi-and-glm-at-scale-wsdraipr3) · Cloudflare · 0 upvotes · 0 comments --- Tags: [#machine-learning](https://daily.dev/tags/machine-learning), [#llm](https://daily.dev/tags/llm), [#cloudflare](https://daily.dev/tags/cloudflare) [View this post on daily.dev](https://daily.dev/posts/introducing-clairic-a-new-system-one-class-of-extraction-models-gzs73ybbr) ```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":"Introducing Clairic, a new System One class of extraction models","url":"https://daily.dev/posts/introducing-clairic-a-new-system-one-class-of-extraction-models-gzs73ybbr","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/introducing-clairic-a-new-system-one-class-of-extraction-models-gzs73ybbr"},"datePublished":"2026-10-19T17:53:18.970Z","dateModified":"2026-10-20T01:45:28.912Z","description":"Cloudflare announced Clairic, a new 'System One' extraction model available as a preview on Workers AI (@cf/cloudflare/clairic). Unlike LLMs, Clairic extracts,...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/5e7302009b855d8f13598258733e1e55?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/5e7302009b855d8f13598258733e1e55?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Cloudflare","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":"Cloudflare","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/38522e1d11354cd6b7af66f9d4316735","url":"https://daily.dev/sources/cloudflare"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/introducing-clairic-a-new-system-one-class-of-extraction-models-gzs73ybbr","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"machine-learning,llm,cloudflare","timeRequired":"PT21M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Cloudflare","item":"https://daily.dev/sources/cloudflare"},{"@type":"ListItem","position":3,"name":"Introducing Clairic, a new System One class of extraction models"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/introducing-clairic-a-new-system-one-class-of-extraction-models-gzs73ybbr#faq","mainEntity":[{"@type":"Question","name":"What is Cloudflare's Clairic model and how is it different from an LLM like GPT or Claude?","acceptedAnswer":{"@type":"Answer","text":"Clairic is a System One extraction model from Cloudflare, available as @cf/cloudflare/clairic on Workers AI, that extracts, computes, estimates, and forecasts typed values (numbers, dates, spans, distributions) from unstructured text or images in a single forward pass with no output tokens. Unlike LLMs, it can't hallucinate text answers, returns confidence scores and p_null probabilities, and is faster and cheaper since it only charges for input tokens. daily.dev surfaces releases like this for teams weighing LLM extraction against purpose-built models."}},{"@type":"Question","name":"How much faster and cheaper is a Clairic plus Clef pipeline compared to using an LLM for extraction and decision steps?","acceptedAnswer":{"@type":"Answer","text":"A pipeline using Clairic for extraction and Clef for decisions, followed by GLM-5.3 for the reply, finished in 19.3 seconds versus 61.7 seconds for an all-LLM version doing every step with GLM-5.3, a 69% time reduction, roughly 3.2 times faster. It also cost 49% less, at $6,032 versus $11,740 per million claims, with Clairic and Clef completing extraction and decision in under a second combined. benchmark numbers like these help developers on daily.dev decide when to swap LLM calls for specialized models."}},{"@type":"Question","name":"What are the current limitations of Cloudflare's Clairic extraction model in its preview release?","acceptedAnswer":{"@type":"Answer","text":"Clairic's reading ability is strong but multi-step arithmetic and aggregating across many table rows are unreliable without a scratchpad, so one- or two-step calculations work better than longer dependent chains. Estimates need supporting evidence like comparables or history, calibrated intervals can run narrow on unfamiliar data, and forecasting is basic, intended for general pattern matching rather than detailed long-horizon predictions. knowing these limits upfront helps developers on daily.dev decide where Clairic fits in a production extraction workflow."}}]} ```