RouterPlex puts Cloudflare Clef on an endpoint its catalog record does not list

RouterPlex puts Cloudflare Clef on an endpoint its catalog record does not list

RouterPlex has priced Cloudflare’s new decision models before Cloudflare’s own promotion ends. In Cloudflare Clef API Pricing: Free for Now, the gateway explains that Clef (27B), Clef-flash (9B) and Clef-omni (30B mixture-of-experts) answer typed questions with a probability per allowed answer and a confidence, not chat text, and that all three went live on RouterPlex on October 18, 2026, free for a limited time. List prices when the promotion ends are $0.24, $0.038 and $0.15 per million input tokens, output free. The call is POST /v1/systemone, the same route as TypeSafe’s Jev, so switching is one word.

The post does the labelling work for us. Cloudflare’s figures — Clef 2.5x faster than Jev at the median and Clef-flash 13x, top score on 7 of 10 decision benchmarks — are flagged as “the vendor’s runs,” and RouterPlex points out that Clef-flash collapses on CLINC150, 66.77 against Jev’s 89.27. Then it runs its own test: one ambiguous support ticket to four models. Clef and Jev returned near coin flips between billing and technical with confidences of 0.23 and 0.34; the smaller Clef models were sure. “If your code gates on confidence, Clef and Jev would have sent this ticket to a person.” That is the useful finding, and RouterPlex is clear that its single cold round trips are “not a latency benchmark.” It also catches a token difference — 262 input tokens on Clef against 371 on Jev for the same request — that makes per-token price comparisons misleading.

The catalog record for RouterPlex has five API pages: the RouterPlex Chat Completions API, RouterPlex Responses API, RouterPlex Messages API, RouterPlex Images API and the RouterPlex Models API. None of them is the decision-model route this post is about. /v1/systemone does not appear in any contract on record, and the agentic access profile counts five operations. The record also describes RouterPlex as fronting chat and image models, which was true before Jev and Clef arrived.

The Kin Score is 46.8, developing, with discoverability at 85.0 — RouterPlex publishes an RFC 9727 API catalog, llms.txt and an MCP server — and contract governance at 18.2. Agent Readiness is 39.8, with error semantics, consent identity and the well-known catalog lit and idempotency and OpenAPI examples dark. A provider this disciplined about labelling numbers ships its newest and most interesting endpoint outside the contract the catalog reads. The fix is one path in the OpenAPI it already publishes.

← Of 284 SLA documents in the catalog, three quote an uptime promise
Solo.io unifies AI and API traffic in agentgateway, and the catalog sees only the portal →