Introducing Clairic, a new System One class of extraction models
Our first one, Clairic, is available hosted on Workers AI as @cf/cloudflare/clairic.We intend to release the model open-weights soon.
Proof3 independent outlets1 community thread
Key points
- Today, we're announcing a preview of a new class of System One-style models: extraction models.
- Clairic is an extraction model that is able to extract, compute, estimate, and forecast values from unstructured inputs — all at a significantly higher accuracy, faster latency, and cheaper cost than traditional LLMs.
- Cloudflare has released an AI model it says pulls data from documents faster and more cheaply than large language models.
- Cloudflare has released a preview of Clairic, a new System One extraction model on Workers AI.
Sources (4)
- [1]Introducing Clairic, a new System One class of extraction modelsCloudflare Blog: AI · Oct 19, 05:48 PM
Our first one, Clairic, is available hosted on Workers AI as @cf/cloudflare/clairic.We intend to release the model open-weights soon.
Today, we're announcing a preview of a new class of System One-style models: extraction models.
- [2]Cloudflare's Clairic pulls data from documents faster and cheaper than LLMsThe Decoder · Oct 20, 12:04 PM
Cloudflare has released an AI model it says pulls data from documents faster and more cheaply than large language models.
- [3]Cloudflare Releases Clairic, a System One Extraction Model on Workers AIMarkTechPost · Oct 20, 07:45 AM
Cloudflare has released a preview of Clairic, a new System One extraction model on Workers AI.
- [4]Introducing Clairic, a new System One class of extraction modelsHacker News: LLM threads (40+ points) · Oct 19, 06:18 PM · same content
Extractive summary: sentences quoted from the sources.
Hype check
Independently evaluated3 independent looks since
- Evaluation · Oct 19, 2026ASTOR: Benchmarking and Improving Autoformulation for Soft-Constraint Optimization with a Formal Grammar
- Evaluation · Oct 19, 2026How to continue a GitHub Copilot app session in VS Code
- Evaluation · Oct 19, 2026Which Depth Estimator for Jetson AGX Orin? A 2026 Benchmark of Monocular and Stereo Methods
- Follow-up · Oct 19, 2026What Language Adds, and at What Cost: Four Controlled Probes of Pathology Vision-Language Models
- Follow-up · Oct 19, 2026Safety Mid-Training: Internalizing LLM Safety as a Foundational Capability
- Follow-up · Oct 19, 2026Bio-inspired Thrust-Assisted Trotting on a legged-Aerial Robot via Real-Time QP-MPC
Before this
- Oct 19, 2026Push-based KV transfer to improve TTFT and TPOT for disaggregated inference in vLLM
- Oct 19, 2026MiniMax M3.1 Flash shows China's AI price race has no floor in sight
- Oct 18, 2026vllm-project/vllm v0.32.0
- Oct 17, 202620 Jev Alternatives: The Decision Models Taking On TypeSafe in 2026
- Oct 17, 2026deepseek-ai/DeepSeek-V4.1-Flash-1017
- Oct 17, 2026Nemotron 3.5 Bolt, Muse Spark Open, and a Fast LFM2.5 24B
