Cloudflare open-sources Clef decision models, up to 13x faster than Jev
dinasaur_404 · x · 2026-10-02
Cloudflare released its first in-house open-source decision models, @cf/cloudflare/clef and clef-flash, on Workers AI with Apache 2.0 weights on Hugging Face. Decision models don't generate text — they read an input state plus typed questions and return probabilities for each allowed answer, giving agents structured decisions (route a ticket, block a request, escalate) with no free-form output to parse. Across 43 benchmark runs, Clef's median latency is 209.3ms (2.5x faster than Typesafe's Jev) and clef-flash hits 38.8ms (13x faster). Running on GPUs across Cloudflare's network, they can sit directly in an agent's hot path before handing off to an LLM. An RL fine-tuning service is also launching with design partners.
More from coding & agent
- Generic API-catalog MCP server vs one tool per job: a context-bloat experiment — MandeepKaur9407 · 2026-11-16
- Berkeley paper: LLMs answer with stale values even when the update is in context — ombre-ai · 2026-11-16
- Effective HTML: 4.2k-star agent skills turn wireframes into self-contained interactive prototypes — tom_doerr · 2026-11-16
- skills-gate: one gate skill fronts 148 agent skills at just 389 tokens — Distinct-Pie2389 · 2026-11-16
- Open-weight Karpo-Decision-31B tops Decision Index at 73.9, one forward pass per answer — CodeByPoonam · 2026-11-16
- shadcn boosts @uisdk: one-shot UI rebuild that's instantly E2E-testable by AI — shadcn · 2026-11-16