AI provider data for your team

A data pilot at $20/month: a weekly provider report and a JSON feed shaped around your stack. Start with the live sample report, then tell us which providers and regions your team needs.

Independent probes across 4 regions; 46 providers in this build. Methodology and measurement limits.

Start with your use case

We agree coverage, delivery and price before starting. No card, no charge, no automatic subscription. We review requests personally.

Data pilot — $20/month

For LLM gateways and teams operating several AI providers. Proposed scope, agreed before payment:

Up to five existing providers across our 4 measured regions.
One filtered JSON report, refreshed daily, plus four weekly summaries of edge response times and failed probes.
One setup handoff and one revision of the agreed output format.

Network checks measure the public API edge. Model response times are included only where we have real inference probes; the sample names every measured model. Account-specific probes, extra regions, real-time paging and an SLA require a separate scope.

Read the working sample · Inspect its JSON. Choose Data for your product below to discuss this pilot. No automatic renewal.

Quick check — $5 one-time

Not ready for a monthly plan? Request a small review of one provider/model and region: a short written summary of available measured response times, recent failures and coverage gaps, with links to the evidence and one practical recommendation.

Based on existing public measurements, not a benchmark of your private prompts. No API keys needed. We confirm coverage before payment; if we cannot answer, we do not charge. Choose Free resources and write "Quick check" in your note below.

Free resources$0Public RSS and JSON are available now. Ask about using them.Personal alertsTarget: $5/monthTell us which outages or model retirements you need to hear about. Delivery is being built; this is an enquiry, not a subscription purchase.Data for your productPilot: $20/monthDiscuss provider and region filters, an agreed JSON format and delivery schedule. Scope and price agreed before any payment.

E-mail (required) Your name (optional) Company or product URL (optional) What do you need help with? (optional)

Optional: topics and providers

Outages Model retirements Prices Providers or models you use

By sending this request you ask us to contact you about it. Your contact details and message are saved on our server and sent to the owner through Gmail to follow up; you are not added to a marketing list. No automated confirmation e-mail is sent yet. To withdraw your request or delete your details, write to contact@llmlatency.dev. Please do not send passwords, API keys or confidential logs.

What could fit your team?

LLM gateways: an external incident record and model-retirement calendar to complement your own routing telemetry.
AI platforms and developer tools: provider/model filters for migration planning and customer-facing notices.
Teams running AI in production: agree which outages matter and how you want to hear about them.

These are pilot options, not a claim of ready-made integrations or a guaranteed SLA. A public synthetic probe does not predict latency or availability for your own account or workload. We first agree on coverage, freshness, format and acceptance criteria.

Try the data before talking to us

Measured incident log · Model-retirement calendar
Incident JSON · Deprecation JSON · RSS

Public data remains CC BY 4.0, including commercial reuse with attribution. A paid pilot would cover agreed preparation, filtering, delivery and support — not permission to use the free data.