OpenClaw
Observed throughTokscale Coding Client Report
OpenClaw is an AI coding product whose local usage records can be imported by Tokscale.
Reality Snapshot
Observed tokens
108.4B-3.3%
over 30 days
Estimated tracked spend
$246K
Invoices may differ
Active developers
48-29
deduplicated over 30 days
Observed messages
1.1M
over 30 days
Most-used model
Claude Opus 545.5%
of observed tokens
Observed Usage
Observed tokensDaily active developers
Models used
Observed 30-day client usage by model
| Model | Tokens | Share | Change |
|---|---|---|---|
| Claude Opus 5Anthropic | 49.3B | 45% | 0% |
| Claude Opus 4.8Anthropic | 29.6B | 27% | 0% |
| GPT-5.6 SolOpenAI | 17.9B | 17% | 0% |
| GPT-5.6 TerraOpenAI | 3.4B | 3.1% | 0% |
| Grok 4.5xAI | 2.9B | 2.6% | 0% |
| Claude Opus 4.6Anthropic | 1.6B | 1.5% | 0% |
| Claude Sonnet 4.6Anthropic | 892.7M | <1% | 0% |
| Claude Fable 5Anthropic | 775.5M | <1% | 0% |
Token composition
Share of observed client tokens
Cache read74.6B69%
Input33.6B31%
Output159.5M<1%
Reasoning00%
Client Economics
Claude Opus 5$135K55%
Claude Opus 4.8$80.2K33%
GPT-5.6 Sol$14.9K6.1%
GPT-5.6 Terra$1.1K<1%
Grok 4.5$1.7K<1%
Claude Opus 4.6$4.7K1.9%
Claude Sonnet 4.6$546<1%
Claude Fable 5$6.9K2.8%
Estimated spend per 1M tracked tokens$2.27
Cache-read share68.9%
Spend basisEstimated from tracked model usage
Estimated tracked spend uses Tokscale model and provider pricing data. Actual invoices may differ.
Client Momentum
-3.3%30-day growthObserved coding client
Compared with the prior non-overlapping 30-day window
Provider footprint
- Anthropic82.1B
- OpenAI21.2B
- xAI2.9B
Client details
Canonical clientopenclaw
TelemetryAggregated Tokscale submissions
Current window2026-07-12 to 2026-08-10
Comparison window2026-06-12 to 2026-07-11
SpendEstimated; invoices may differ
About this client
OpenClaw is an AI coding product whose local usage records can be imported by Tokscale.
Links & sources
TOKSCALE INSIGHT
OpenClaw recorded 108.4B coding tokens across 48 developers during the latest 30-day Tokscale window.
Evidence confidenceStable observed sample