{"slug": "cloudflare-clairic-open-weights-updated-27b-and-new-9b-clairic-flash", "title": "Cloudflare Clairic Open Weights: Updated 27B and New 9B Clairic-Flash", "summary": "Cloudflare released open-weight Clairic (27B) and Clairic-flash (9B) extraction models on November 18, 2026, hosted on Workers AI and published on Hugging Face, graduating Clairic from preview and adding a 9B flash tier. Cloudflare reports both new models scored 1,263/1,263 correct fields on internal tracked assertions versus 1,033/1,263 for the preview, and 576/576 on synthetic receipt fields versus 369/576 for preview, with image extraction errors dropping from 207 to zero on an internal mixed-input set. Cloudflare measured p50/p95 full HTTP response times of 358/736 ms for Clairic and 155/328 ms for Clairic-flash on fresh text fields, and 453/976 ms and 305/605 ms respectively on fresh image fields, charging input tokens only with no separate output-token meter.", "body_md": "**Cloudflare released open-weight [Clairic](https://blog.cloudflare.com/open-clairic/?ref=explainx) and [Clairic-flash](https://blog.cloudflare.com/open-clairic/?ref=explainx) on November 18, 2026**, hosted on **Workers AI** and published on **Hugging Face**, with the **Workers AI Superintelligence team** branding the launch. The update **graduates Clairic from preview**, adds a **9B flash tier**, and **re-architects extraction** for **PDFs, images, tables, and dates**.\n\nIf you already use **System One** models to **shrink chat bills**, this is the **extraction-specific** release that pairs with **[Clef](https://explainx.ai/blog/clairic-vs-jev-vs-openai-decisions-api-vs-clef-vs-microsoft-decision-1-comparison-2026)** for **routing** and **decisions**.\n\n## What did Cloudflare announce?\n\nThe [November 18 blog post](https://blog.cloudflare.com/open-clairic/?ref=explainx) names two deliverables:\n\n| Model | Role | Workers AI ID | \n|---|---|---|\n| **Clairic (27B)** | Flagship typed extraction | `@cf/cloudflare/clairic` | \n| **Clairic-flash (9B)** | Faster, cheaper extraction | `@cf/cloudflare/clairic-flash` | \n\nBoth ship as **open weights** and **managed endpoints**. Cloudflare says the mission is **purpose-built small models** that **augment expensive LLM usage** rather than **replace frontier chat models entirely**.\n\n## What is System One extraction?\n\n**System One** is Cloudflare's label for **narrow models** that return **typed values** — numbers, dates, enums, spans — instead of **free-form text**. Our earlier **[claimed-vs-verified Clairic preview piece](https://explainx.ai/blog/cloudflare-clairic-system-one-extraction-model-workers-ai-claimed-vs-verified-2026)** walked the preview API.\n\nThe **November 18** release keeps the **schema-first contract**:\n\n- **Confidence scores** and**character spans** per field\n- **Structured fields** you define in the request\n- **No separate output-token meter** ; Cloudflare**charges input tokens only**\n\nThat pricing shape matters when you **extract hundreds of fields** from **invoices, leases, or support tickets**.\n\n## Architecture changes since preview\n\nCloudflare says **preview feedback** pushed a **full redesign**:\n\n- **Parallel slot autoregressive generation** for ordinary text and number fields, so each field**stops at its own end marker** while the batch advances together\n- **Special heads** for**estimates, forecasts, durations, calculated dates, and arithmetic**\n- **PDF and image ingestion** beyond the**receipt photo** demo from preview\n\nOn **internal tracked assertions**, Cloudflare reports **1,263/1,263** correct fields for both new models versus **1,033/1,263** for preview. On **synthetic receipt fields**, it reports **576/576** for Clairic and Clairic-flash versus **369/576** for preview.\n\nTreat those tables as **vendor evals**, not independent benchmarks.\n\nCloudflare Clairic internal eval scores for text image PDF and receipt extraction\n*Image: [Cloudflare](https://blog.cloudflare.com/open-clairic/?ref=explainx)*\n\n## Receipt and image extraction example\n\nCloudflare Clairic open-weight model receipt extraction comparison versus preview\n*Image: [Cloudflare](https://blog.cloudflare.com/open-clairic/?ref=explainx)*\n\nCloudflare's side-by-side **coffee shop receipt** shows **preview returning null** for the merchant name while **Clairic** returns **Coffee House** with **all five fields** filled. Image extraction error counts moved from **207 errors to zero** on an internal mixed-input set, per the blog.\n\nFor **operators**, the lesson is **validate on your document distribution**. Receipt OCR is **not** the same as **scanned contracts** or **noisy mobile photos**.\n\n## Latency: is Clairic fast enough for online flows?\n\nCloudflare measured **full HTTP response time** on a **self-hosted server** and published **p50 / p95** in milliseconds:\n\n| Eval set | Clairic p50 / p95 | Clairic-flash p50 / p95 | \n|---|---|---|\n| Fresh text fields | 358 / 736 ms | 155 / 328 ms | \n| Fresh image fields | 453 / 976 ms | 305 / 605 ms | \n| Commercial leases | 691 / 803 ms | 379 / 519 ms | \n| Loan pricing | 395 / 407 ms | 188 / 192 ms | \n\n**Clairic-flash** targets **sub-second** responses on several sets **below 400 ms p50**. **Full Clairic** still lands **under one second** on many rows, which Cloudflare contrasts with **typical chat completions** on similar workloads.\n\n## Pricing and the December 1 promo\n\nCloudflare moved the **9B preview endpoint** to **Clairic-flash** and placed the **27B model** on the main **Clairic** ID.\n\n| Model | List input price after promo | Promo through Dec 1, 2026 | \n|---|---|---|\n| [Clairic 27B](https://developers.cloudflare.com/workers-ai/models/clairic?ref=explainx) | $0.24 per million input tokens | $0.10 | \n| Clairic-flash 9B | $0.10 per million input tokens | $0.10 | \n\nIf you pinned **preview Clairic at $0.10**, Cloudflare says **the endpoint upgrades in place** to **better extraction** without a **schema migration**. **Cost-sensitive** jobs should **test Clairic-flash** before **December 1**, when **27B returns to $0.24**.\n\nCompare this shape to **[OpenAI Decisions API](https://explainx.ai/blog/clairic-vs-jev-vs-openai-decisions-api-vs-clef-vs-microsoft-decision-1-comparison-2026)** and **[Microsoft Decision-1.1](https://explainx.ai/blog/microsoft-ignite-mai-code-2-seven-models-november-2026)** when you **route** versus **extract**.\n\n## Open weights on Hugging Face\n\nCloudflare follows **Clef's open-weight playbook**: publish weights for **research and fine-tuning**, keep **managed inference** on **Workers AI**.\n\nReasons builders download weights:\n\n- **Air-gapped** document pipelines\n- **Fine-tunes** on**industry-specific forms**\n- **Latency experiments** on**your own GPUs**\n\nReasons to stay on **Workers AI**:\n\n- **No kernel tuning**\n- **Billing integration** with**Workers** and**R2**\n- **Automatic model router** updates\n\nCheck **license files** on Hugging Face before **redistributing** derivatives.\n\n## Competitive context: extraction is crowded\n\nCloudflare compares against **Liquid AI D1.2 with Apple Vision OCR** and **Scaledown Extract** in its tables. It admits **no public extraction leaderboard** exists yet.\n\nThat honesty matters. **Extraction** is becoming a **commodity layer** as **agents** ingest **email, PDFs, and tickets**. **[Jev](https://explainx.ai/blog/clairic-vs-jev-vs-openai-decisions-api-vs-clef-vs-microsoft-decision-1-comparison-2026)** and **Microsoft's Decision models** sit adjacent but **optimize decisions**, not **field filling**.\n\nIf you run **agents on Cloudflare**, validate every extracted field before it triggers a tool call when **untrusted documents** enter the loop.\n\n## How to try Clairic this week\n\n1. Open **Workers AI** docs for**`@cf/cloudflare/clairic`** and**`@cf/cloudflare/clairic-flash`** .\n2. Port **one production form** with**five to ten fields** and compare**preview-era** accuracy.\n3. **Load-test****Clairic-flash** if**p95 under 400 ms** is a**hard SLA** .\n4. **Download Hugging Face weights** only if**self-host TCO** beats**$0.10/M** promo math on**your volume** .\n\n## Workers AI Superintelligence team\n\nCloudflare reframes its frontier research group as the **Workers AI Superintelligence team**, responsible for **Clef**, **Clairic**, and **unannounced models**.\n\nFor **edge developers**, the takeaway is **Cloudflare building models**, not only **hosting others' checkpoints**. That affects **router defaults** and **which open weights** get **first-class** Workers integrations.\n\n## PDFs, leases, and multi-page line items\n\nCloudflare's **PDF support** is the **largest user-visible upgrade** since preview. You can request **supplier, line items, and totals** when fields **span pages**. The cap is **four images or PDF pages per request**, which pushes designers toward **chunking** very large packets.\n\nFor **commercial leases**, Cloudflare's latency table shows **691 ms p50** on full Clairic and **379 ms** on flash. **Loan pricing** rows are **tighter** — **395 ms** versus **188 ms** — suggesting **numeric tables** compress well in the **9B model**.\n\nIf you migrate from **general chat extraction**, benchmark **three document types** you actually ship: **invoices**, **contracts**, and **support exports**. Cloudflare's **tracked customer assertions** row is the closest proxy to **messy production**.\n\n## Self-hosted vs Workers AI: a cost sketch\n\nOpen weights enable **private cloud** deployment, but **total cost** is **GPU hours + engineer time**, not only **$/M tokens**.\n\nA rough decision tree:\n\n- **Under a few million fields per month** — stay on**Workers AI** during the**$0.10 promo** , especially if you already pay for**Workers** and**R2** .\n- **Strict data residency** — download**Hugging Face** weights and run on**your VPC** , accepting**higher p95** until you tune kernels.\n- **Burst-heavy batch** — compare**Clairic-flash****throughput** against**batching** on**your own vLLM** cluster.\n\nCloudflare's **input-only pricing** simplifies **math**: you do not multiply by **output tokens** when **schemas return large JSON**.\n\n## Pairing Clairic with agents and routers\n\nMost production stacks should **not** send **every user message** to **Clairic**. Typical layering:\n\n1. **Clef or Decision model** — classify intent and pick tools\n2. **Clairic** — fill**typed slots** from**attachments**\n3. **Frontier chat model** — synthesize**narrative answers** only when needed\n\nThat pattern mirrors **[Microsoft IQ plugins](https://explainx.ai/blog/microsoft-ignite-2026-complete-coverage-every-announcement-november-2026)** and **[OpenAI Decisions API](https://explainx.ai/blog/clairic-vs-jev-vs-openai-decisions-api-vs-clef-vs-microsoft-decision-1-comparison-2026)** stories from the same month.\n\n## Eval tables: how to read Cloudflare's numbers\n\nCloudflare publishes **fresh text**, **mixed-input**, **customer assertions**, and **synthetic receipt** rows. **Clairic and Clairic-flash tie** on several **perfect scores**, while **preview lags**.\n\nWhen vendors compare against **Liquid D1.2 + OCR** and **Scaledown Extract**, note **different OCR pipelines**. **Apple Vision** on **D1.2** is **not** identical to **Clairic's native image heads**.\n\nUntil a **neutral leaderboard** exists, treat comparisons as **directional** and run **your own 200-field golden set**.\n\n## Security and untrusted documents\n\nExtraction models **read attacker-controlled PDFs**. That is **indirect injection** surface when an **agent** auto-invokes Clairic on **email attachments**.\n\nCombine **schema validation** with **[agent guardrails](https://explainx.ai/blog/promptarmor-frontierbreach-hidden-instructions-agent-benchmark-november-2026)** when **untrusted files** enter the loop. **Typed JSON** does not stop **malicious field content** from flowing downstream.\n\n## What developers are saying\n\nThe [Hacker News thread](https://news.ycombinator.com/item?id=50552988&ref=explainx) was **quiet at launch**, with **few hands-on reports** yet. When **builders post latency numbers** from **real PDFs**, treat them as **complements** to Cloudflare's **self-hosted** measurements.\n\n## Related reading\n\n## Migration checklist from Clairic preview\n\nIf you integrated **`@cf/cloudflare/clairic`** during preview, Cloudflare says **weights and schemas roll forward** without renaming your **Workers binding**. Still run this checklist before **December 1**:\n\n- **Re-benchmark** ten production documents; preview**null fields** may disappear.\n- **Switch latency-sensitive routes** to**Clairic-flash** and compare**p95** against your**SLA** .\n- **Pin Hugging Face revisions** if you**self-host** , because**open-weight tags** can move independently of**Workers AI** defaults.\n- **Update cost dashboards** when**27B returns to $0.24** per million input tokens.\n\nDocument **schema versions** in your repo so **finance** can reconcile **November promo** spend against **December list prices**.\n\n*Pricing and model IDs reflect Cloudflare's November 18, 2026 announcement. Promo pricing ends December 1, 2026 unless Cloudflare extends it.*", "url": "https://wpnews.pro/news/cloudflare-clairic-open-weights-updated-27b-and-new-9b-clairic-flash", "canonical_source": "https://explainx.ai/blog/cloudflare-clairic-clairic-flash-open-weight-workers-ai-november-2026?utm_source=rss&utm_medium=feed&utm_campaign=blog", "published_at": "2026-11-18 00:00:00+00:00", "updated_at": "2026-11-19 00:15:58.767557+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "ai-products", "ai-tools", "ai-infrastructure"], "entities": ["Cloudflare", "Clairic", "Clairic-flash", "Workers AI", "Hugging Face", "Workers AI Superintelligence team", "Clef"], "also_reported_by": [], "takeaway": "Clairic-flash (9B) returns fresh text fields at 155 ms p50 versus 358 ms for the 27B Clairic, and both models bill input tokens only with no separate output-token meter.", "recommendation": null, "alternates": {"html": "https://wpnews.pro/news/cloudflare-clairic-open-weights-updated-27b-and-new-9b-clairic-flash", "markdown": "https://wpnews.pro/news/cloudflare-clairic-open-weights-updated-27b-and-new-9b-clairic-flash.md", "text": "https://wpnews.pro/news/cloudflare-clairic-open-weights-updated-27b-and-new-9b-clairic-flash.txt", "jsonld": "https://wpnews.pro/news/cloudflare-clairic-open-weights-updated-27b-and-new-9b-clairic-flash.jsonld"}}