State of AI Coding Spend 2026

Everyone has an opinion about what AI coding agents cost. We have receipts. Developers on the Viberank leaderboard measure their own usage locally with ccusage — real Claude Code, Codex, and Gemini CLI logs, not survey answers — and submit it publicly. This report covers everything on the board as of June 2026.

The median serious user has burned $1,285

Lifetime usage value per developer, at API-equivalent prices:

The top user has consumed $56,694 of compute value across 81 billion tokens. The mean ($2,904) sits at more than twice the median, because AI coding spend is a textbook power law:

• The top 1% of users account for 14% of all spend
• The top 10% account for 51%
• The top 25% account for 74%

Daily burn: $29 is the new normal, $200+ days are routine

Across all 29,000 tracked coding days:

11% of all tracked days exceeded $200 — an entire Claude Max monthly subscription's worth of compute, consumed in one day. The single biggest day we've recorded: $3,820. And the median submitter is active 26 days per submission window — this is daily-driver usage, not weekend tinkering.

Power users have outgrown their subscription price 5–10×

Normalizing each developer's usage to a monthly rate: roughly half run at $1,000+/month in API-equivalent value, and ~73% run above $400/month. Most of them pay $100–200/month flat for Claude Max. The arbitrage is stark — the median heavy user extracts 5–10× the sticker price of their plan in raw compute. Flat-rate pricing made always-on agentic coding economically rational, and the usage curves show developers responded exactly as you'd expect.

AI coding is 95% rereading, 0.2% writing

The token mix across the entire dataset is the most lopsided stat in this report:

For every token an agent actually writes, it re-reads ~406 tokens of cached context. The 2.5 trillion token headline is really ~5.9B tokens of generated work product riding on a mountain of context re-reads. This is why prompt caching, not model price, is the real economic engine of agentic coding — and why long sessions in big repos get expensive even when output is small.

The always-on cohort exists

Seven developers have logged 200+ active days, and the longest consecutive-day streak on the board is 238 days straight — eight months without missing a single day. The top of the leaderboard isn't people who type fast; it's people who've turned agents into infrastructure that runs whether they're at the keyboard or not.

Weekends barely slow anyone down: Saturdays and Sundays are 24% of active days (a uniform week would be 28.6%), and a weekend day burns the same ~$82–84 as a Tuesday. When your agent does the typing, "logging off" is a softer concept.

Model mix: Opus is the workhorse, not the treat

Nine in ten developers run Opus — the most expensive tier is now the default for serious work, not a special occasion. Haiku's 64% tells the quieter story: cheap models doing subagent and utility passes inside bigger workflows.

Multi-agent developers are 9% of the board — and most of the top 10

A note on coverage: viberank started as a Claude Code leaderboard and only recently opened to every agent ccusage tracks — Codex, Gemini CLI, Copilot, OpenCode and friends. So this dataset is still overwhelmingly Claude Code, and the cross-tool numbers should be read as early signal, not market share.

That said: about 9% of the board already reports usage from more than one coding agent, and they're heavily over-represented at the very top — several top-10 spenders run three or more agents in parallel. If you want a preview of the median 2027 workflow, look at the current multi-tool tail — see our Codex vs Claude Code vs Gemini CLI comparison.

If you use Codex, Gemini CLI, or anything else alongside (or instead of) Claude Code, your submission makes the next edition of this report meaningfully better — one command covers every tool ccusage detects:

npx viberank-cli

Usage is at an all-time high

Usage value logged per calendar month across all submissions:

After the launch spike (Jul–Aug 2025) and an autumn lull, tracked usage has climbed for six straight months. Spring 2026 is running at launch-hype levels — except now it's steady-state behavior, not novelty.

Who these developers are

Cross-referencing public GitHub data for the board: the median account is ~9 years old, a third have 50+ public repositories, and spend rises with experience — developers with 13+ year-old GitHub accounts have the highest median usage. Geographically (by site traffic): US ~31%, South Korea ~13% — the strongest per-capita AI-coding adoption signal we see anywhere — then the UK, India, and Canada. AI-assisted coding at this intensity is a senior-engineer phenomenon, not a shortcut for beginners.

Methodology & honest caveats

• Source: developers run ccusage locally against their own agent logs and submit the result. GitHub-OAuth submissions are verified (blue check); CLI submissions are self-reported.
• "Spend" means API-equivalent value — what the tokens would cost at list API prices. Most users actually pay a flat subscription, so real out-of-pocket cost is usually far lower.
• Selection bias is real: people who submit to a usage leaderboard skew heavy. Read these numbers as benchmarks for serious users, not population averages.
• Stats use each developer's largest submission to avoid double-counting overlapping date ranges.
• Coverage skews Claude Code: multi-tool submissions only opened recently, so other agents are underrepresented relative to their real-world usage.

Where do you rank?

Every number in this report comes from developers who took thirty seconds to submit. One command measures your real usage — across Claude Code, Codex, Gemini CLI, and every other agent ccusage tracks — and puts you on the board:

npx viberank-cli

Prefer not to install anything? Sign in with GitHub on viberank.app and upload your cc.json for a verified blue check.

Curious what drives the bill — or how to shrink it? Read how much Claude Code costs and 9 ways to cut your AI coding bill.