You.com routes 12 open-weight models behind one enterprise API

You.com launched Inference Router on October 22, and the builder debate since has been about whether its two enterprise promises survive contact with audits. The service routes 12 open-weight models through one OpenAI-compatible endpoint with a claimed 99.9% uptime SLA and zero data retention; the announcement post drew roughly 395,000 views in a day.

Key facts

• Business Wire confirms the launch, 12 models, a 99.9% uptime SLA, zero data retention, and US hosting
• The announcement post on X drew about 395,000 views and 166 likes in the first day, per the post's own counters
• You.com publishes Inference Router usage documentation with an OpenAI-compatible request format
• The uptime SLA and zero-retention commitments are provider claims; no independent audit or attestation has been published yet

What's driving the conversation

Richard Socher's team announced the router on X and amplification from @hasantoxr carried it past 395,000 views. The replies split into two camps. One treats SLA-backed open-weight routing as the missing enterprise on-ramp: a procurement team can now put an open model behind a contract with credits attached instead of defending self-hosting to a risk committee. The other treats the claims as the story, pointing out that "zero retention" from a hosted router says nothing about what the upstream inference provider logs.

The launch also surfaced a discovery gap worth naming. Business Wire carried the release on October 22 and trade writeups followed quickly, but the story missed our own RSS and social sweeps for two passes before surfacing late on October 23. Provider-launch wire releases clearly need a standing check in any monitoring rotation.

The substance

What is verified: the launch happened, the service routes 12 open-weight models, the endpoint is OpenAI-compatible, hosting is US-based, and the documentation is live. What is claimed: the 99.9% uptime SLA and zero data retention. A claimed SLA becomes real when the published terms specify credits and exclusions; a zero-retention claim becomes verified only with a retention policy document or third-party attestation, as our piece on zero-retention AI APIs argues. Neither exists at publication time. The catalog reportedly includes Kimi-class frontier open models, which would put today's Kimi K3.1 Pro API launch and this router in the same week's story.

Why builders are watching

Routers sit exactly where the managed-vs-self-hosted comparison gets decided, and an SLA on open weights narrows the reliability argument for hyperscaler closed APIs. The open question for You.com is whether the enterprise commitments come with the paperwork enterprises require. If the SLA terms and a retention attestation land, Inference Router becomes a serious option for compliance-sensitive workloads on open models; if not, it is one more endpoint in a crowded market. Watch the published SLA terms and any audit on the You.com docs page.