A collective experiment

What is AI actually costing all of us?

Anonymized, aggregate usage across everyone who opted in, over the last 30 days. No names, no projects, no single person's numbers. The more people who contribute, the sharper the shared picture gets.

Updated 4 minutes ago. Refreshes automatically.

Community spend (30d)

$447.58

$17.90/day avg

Tokens processed

52.2B

882 events

Tokens per dollar

116.7M

incl. cached reads

Contributors

5

peak $40.81 on Jul 17

Right now the community runs mostly on Anthropic, via Claude Code , for Coding. Add your stack and these break wide open.

Daily community spend

Estimated spend per day across all contributors.

$0.00 $15.2 $30.4 $45.5 Jul 1: $1.35 Jul 2: $16.9 Jul 3: $18.2 Jul 5: $17.6 Jul 7: $20.9 Jul 8: $12.8 Jul 9: $4.87 Jul 11: $29.5 Jul 12: $20.3 Jul 13: $11.3 Jul 15: $8.49 Jul 16: $33.2 Jul 17: $40.8 Jul 18: $15.9 Jul 19: $24.3 Jul 21: $5.59 Jul 22: $17.8 Jul 23: $13.1 Jul 24: $21.6 Jul 26: $7.99 Jul 27: $11.1 Jul 28: $12.2 Jul 29: $9.60 Jul 30: $34.9 Jul 31: $37.3 Jul 1 Jul 31

What a million tokens lists for today

$4.64

Frontier

$0.70

Budget

See the AI Cost Index →

Outcome signals

No outcome data yet. Send success, latency_ms, and quality_score with your events to unlock community-wide reliability, speed, and quality. See the API.

By provider

  • Anthropic · 100% $447.58

    52.2B tokens · 882 events

By model

  • Claude Opus 4 8 · 7% $30.20

    4.1B tokens · 12 events

  • Claude Sonnet 5 · 5% $23.50

    922.0M tokens · 44 events

  • Claude Opus 5 · 5% $20.37

    2.2B tokens · 13 events

By platform

  • Claude Code · 100% $447.58

    52.2B tokens · 882 events

By use case

  • Coding · 100% $447.58

    52.2B tokens · 882 events

Every event carries more than a number

We capture each of these per usage event, then aggregate them anonymously. The more of you who send the optional signals, the richer every chart above gets.

Provider Model Platform Use case Input tokens Output tokens Cached tokens Reasoning tokens Cost (actual or estimated) Latency Success Quality score

Everything here is aggregate and anonymized. We only show combined totals across people who chose to contribute, and we hide any provider, model, platform, or use case shared by fewer than 3 contributors, so no individual can be identified. We never publish your name, your projects, your working directories, or your own numbers. You can leave the community anytime from your settings, and your data drops out of the next refresh.