Cloudflare Open Sources Decision Models for AI Agents
Cloudflare open-sources decision models for AI agents—9B and 27B parameters designed to choose, not generate.
Why it matters
Cloudflare releases specialized open-weight models optimized for classification and bounded decision-making rather than text generation. Relevant for practitioners building agent scaffolding and enterprises seeking alternative model architectures for constrained reasoning tasks.
The key facts
13 to knowCloudflare released Clef: open-weight decision models (9B and 27B parameters)
Models designed to choose between predefined options rather than generate text
Platform included for adapting models to specific decision-making tasks
Announced during Cloudflare Birthday Week (October 2026)
Open-source release with adaptation framework
Architecture distinction: classification/selection vs. generation—relevant for agent control and reliability
Cloudflare released Clef during Birthday Week 2026
Two open-weight models: 9B and 27B parameters
Models designed for classification/decision-making, not text generation
Platform provided for adapting models to specific tasks
Open-source release—weights publicly available
No pricing, performance benchmarks, or competitive positioning disclosed
No details on training data, safety testing, or production deployment examples
The story so far
Earlier coverage of this storyline
- Cloudflare Releases Clef and Clef-flash: Open-Weight Decision Models That Return Typed Probabilities Instead of TextMarkTechPost
- AWS Strands Labs Releases Strands Decider 2B: An Open Source Decision Model That Picks Options in About 115 msMarkTechPost
- Multimodal open d1 decision models for the edgeHugging Face Blog
- This story
Go to the source
InfoQ AI/MLinfoq.com
Publisher excerpt: During its recent “Birthday Week”, Cloudflare announced Clef, a set of open-weight AI models designed to choose between predefined options rather than generate text. Cloudflare released 9B- and 27B-parameter models, along with a platform for adapting them to specific decision-making tasks. By…