As enterprise AI coding spend accelerates into the billions, cost visibility without optimization is just a bigger invoice.

Claude Code and Codex are changing how software gets built. Engineering teams are moving fast — using AI to write code, fix bugs, invoke tools, and automate workflows that used to take days. For the teams pulling ahead, these tools aren't a productivity experiment. They are core infrastructure. Comet has a unique view into how that adoption is scaling — and what it's costing.

As adoption scales, most engineering leaders can't answer basic questions: How much are we spending per engineer and per team? Is that spend going toward new features and bug fixes or something else entirely? Are we burning tokens on MCPs, skills, and plugins we don't need? Those questions are turning into real budget conversations — and the data to answer them doesn't exist yet. Until now.

Today, Comet announced coding agent cost tracking in Opik, its AI observability and evaluation platform — giving engineering leaders answers their bills can’t. Per engineer, per team, per task. In real time, before the invoice arrives.

Opik goes beyond observability to automatically slash your coding agent costs. By eliminating unused skills, keeping idle MCPs out of your context, and fixing misconfigured compaction strategies, Opik optimizes your spend behind the scenes. One enterprise customer reduced AI spend by millions annually with zero restrictions on how their developers use AI.

“Claude Code and Codex are genuinely transformational tools,” said Gideon Mendels, Co-founder and CEO of Comet. “But most engineering leaders have no idea how their developers have them configured – which MCPs are loaded, which model is running by default, whether any of it maps to real outcomes. Opik answers the questions every VP of Eng should be asking right now — and gives them the tools to do something about it.”

Built for Teams Building Seriously With Claude Code and Codex

• Spend per engineer and per team — see exactly how much Claude Code and Codex are costing across every developer and project in a single real-time view, updated continuously so surprises don’t wait until month end
• Attribute spend to real work — understand whether AI spend is going toward new features, bug fixes, or tasks that didn’t need it — giving engineering leaders the data to justify spend and redirect it where it matters
• Eliminate token waste — see which MCPs, skills, and plugins are loaded to context for every task, identify what’s unnecessary, and standardize configuration across the team before wasted tokens compound into a real budget problem

About Comet

Comet builds tools for AI engineering teams moving from experiment to production. Founded in 2017, Comet is trusted by thousands of teams building with LLMs. Opik, Comet’s AI observability and evaluation platform, is available at https://www.comet.com/site/products/opik/.

View source version on businesswire.com: https://www.businesswire.com/news/home/20260609844223/en/

Most engineering leaders have no idea how their developers have Claude Code and Codex configured – which MCPs are loaded, which models and skills are running. Opik Cost Intelligence answers these questions and provides tools to optimize coding agent spend.