Skip to content
NewFinOps, Kubernetes and AI, on one platform
Book a demo

GenAI cost control

Keep GenAI spend in step with the value it creates

Measure effective cost per 1M tokens across every model and cloud, move workloads to the right model and capacity, and stop any team from overrunning its budget.

Book a demoPlatform overview

What UnitVane brings to this job

What a million tokens really costs you, by model and team

Tokenomics

Billing joined to token telemetry for every deployment: input, output and cached tokens, PayGo, PTU and reservation cost, and an all-in cost per 1M tokens that you can compare across models and clouds.

Explore Tokenomics

The problem

Why it is hard today

Cost per 1M tokens, budgets per team, and PTU that pays for itself

  • Three pricing models for one model:

    Pay-as-you-go, PTU hours and reservations make cost per token hard to compare.

  • Expensive models doing simple work:

    Summaries and rewrites run on the largest model because it was the default.

  • Budgets that only warn:

    By the time an alert arrives, the month's spend has already happened.

How UnitVane handles it

What you get

Effective cost per 1M

All-in, by model, team and cloud.

Model-mix savings

Cheaper models for the same tier of work.

PTU right-sizing

Deployments sized from per-minute utilization.

Billing correction

Hourly PTU separated from pay-as-you-go.

Budgets and alerts

By product, team, tenant or deployment.

Hard stops

Gateway keys stop at 100% of budget.

Outcomes

What changes for your team

One number to manage

Cost per 1M tokens, trended per model and team.

Budgets that hold

Enforced at the gateway, not just reported.

Your first weeks

What you see, and when

From the first provider connection to capacity you can size and commit.

  1. 01

    When you connect a provider

    Billing and token telemetry are read for each deployment, read-only. Effective cost per 1M tokens appears for every period whose billing has arrived, with nothing routed through the gateway.

  2. 02

    From the first gateway request

    Each routed request is metered as it is served: tokens, route, target, latency and cost, attributed to the team key that sent it.

  3. 03

    After 7 days of utilization

    Provisioned deployments have enough per-minute data to be sized to the percentile you choose.

  4. 04

    After 14 days

    Deployments count toward reservation coverage, so over-reserved and uncovered capacity are reported per pool.

UnitVane works with your technology

Every major cloud, any Kubernetes, and the tools and model providers your teams already use.

All integrations

Clouds and Kubernetes: AWS, Microsoft Azure, Google Cloud, Oracle Cloud, Kubernetes, Red Hat OpenShift, Rancher, VMware Tanzu, k3s

Your tools and model providers: Helm, Terraform, Argo CD, Flux, Prometheus, OpenTelemetry, Okta, Slack, Microsoft Teams, OpenAI, Anthropic, Mistral AI

Your questions answered

Keep reading

Expert guides and related solutions

Article

What a million tokens really costs

The price list says one number per million tokens. Your bill says another. Provisioned capacity billed by the hour, cached tokens, retries and output ratios decide the real cost, and whether a reservation pays off.

Article

Idle GPUs: the most expensive capacity you don't use

GPU spend is growing fastest of any compute, and published research from Microsoft and Alibaba shows how little of it does work. Why GPUs sit idle in Kubernetes, how to measure it, and what sharing, sizing and pricing change.

Solution by use case

Showback and chargeback

A bill for every team, idle and shared included

Solution by use case

Kubernetes rightsizing

Requests and nodes sized from measured usage

Solution by use case

Idle and waste cleanup

Abandoned workloads and orphaned volumes, found and priced

All resources

See your AI spend per team, model and request

A 30-minute walkthrough on your own Azure OpenAI, Bedrock or Vertex AI usage.

Book a demoGateway quickstart