Shipped: Log and track Insights from your AI assistant

CloudZero continues to ship new features and enhancements at a fast pace. We're publishing these product update posts to keep you up to date on our velocity and how we're adding customer value.

AI cost per query: how to calculate and optimize it

Providers disclose what a query costs them; none tell you what it costs you. AI cost per query is your fully loaded serving cost per request — the base unit of AI unit economics. Here's how to calculate it, what moves it (a 22x model spread), and how to track it.

How much does it cost to train an AI model? Frontier figures, graded [2026]

Training an AI model can cost $900 or several hundred million, depending entirely on what training means. Every famous frontier figure, graded by provenance, plus what a from-scratch run actually costs.

Veo 3 pricing in 2026: every tier, API rate, and what AI video actually costs

With the Sora API shut down, Google Veo is where most AI video budgets are heading. Here's what every Veo 3.1 tier and Google AI plan costs, why real spend runs well above the rate card, and how to measure video cost per unit of value.

AI cost allocation: how to attribute AI spend by team, product, and customer

AI spend arrives untagged, spans vendors, and pools in shared infrastructure. AI cost allocation attributes every dollar to the team, product, or customer that generated it — here are the four methods that work and how to handle the shared and untagged parts.

We stopped asking an LLM how much its own work would cost

How CloudZero's model-routing agent stopped asking an LLM to guess a token budget and switched to six bounded signals plus a deterministic formula: why averaging capped accuracy, why some signals dilute the ensemble, and where the v1.0.0 formula still needs calibration.

Build vs. buy: should you build your own AI cost management tooling?

Building your own AI cost tooling is a multi-quarter platform project with a permanent maintenance tail. Here is the four-question framework, a full TCO model in engineer-years, and the honest case for when building actually wins.

Kling AI pricing in 2026: plans, credit costs, API packages, and the spend no invoice shows

A finance guide to Kling AI pricing in 2026: the subscription tiers and why they're hard to pin down, how per-second credits actually work, what the separate API packages cost, how Kling compares with Veo and Runway, and how to track spend that never hits a cloud bill.

ElevenLabs pricing in 2026: plans, credits, and agent costs

A finance guide to ElevenLabs pricing in 2026: the six subscription tiers, how credits translate into minutes, what the API and voice agents actually cost per unit, and the three meters, ElevenLabs, LLM, and telephony, behind every voice-agent call.

Token-based pricing: how AI usage billing works (2026)

A finance guide to token-based AI pricing in 2026: how input and output tokens turn into dollars, what the major models charge per million tokens, the discounts and surcharges that move the effective rate, and how to forecast and govern the bill.