Cloudflare has released Clef-Flash, a 9 billion parameter decision model designed for speed and efficiency, built upon Qwen3.5-9B. Unlike traditional chat models, Clef-Flash processes input states and typed questions in a single forward pass, returning probabilities for predefined answers without generating free-form text. This makes it suitable for tasks like support ticket triage, where it can quickly classify urgency and ownership. Cloudflare also offers a larger, precision-tuned 27B parameter model named Clef, which includes vision capabilities for image classification. AI
IMPACT Introduces a new class of decision-focused models that could streamline AI applications by bypassing traditional text generation for classification tasks.
RANK_REASON Release of a new, specialized LLM by a major tech company, with details on its architecture and performance benchmarks. [lever_c_demoted from research: ic=1 ai=1.0]
- Banking77
- Clef-Flash
- CLINC150+OOS
- Cloudflare
- Hugging Face
- Jevíčko
- llama.cpp
- MLX
- Ollama
- Qwen3.5:9b
- SystemOne
- When2Call
- Workers AI
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →