CATCH #744
DETECTED AUG 15, 2026 15:31 UTC
THE CATCHscore 7/10
EXHIBIT 744 · URGENT
Aug 15, 2026 15:31 UTC
web app
https://platform.claude.com
Anthropic's Claude platform deployed new frontend strings suggesting an upcoming feature to break down output tokens by category, including internal reasoning tokens.
READY TO POST:
looks like claude is adding a feature to show a breakdown of output tokens, including those used for internal reasoning. good for observability!
anthropic's platform bundle suggests they're working on exposing internal reasoning tokens as a separate category in output usage. interesting for cost tracking.
EVIDENCE — the receipts
⚠️ FROM WEB-APP BUNDLE — may include vendored-dependency strings, verify before posting.
Anthropic Console (https://platform.claude.com) deployed new frontend code [0e7990e3721d1b2d -> aca8e3c38b49193a]: +209 strings, -595 strings, 2 interesting
[feature_flag] (2)
+ Breakdown of output tokens by category.
`output_tokens` remains the inclusive, authoritative total used for billing.
This object provides a read-only decomposition for observability — for example,
how many of the billed output tokens were spent on internal reasoning that may
have been summarized be
+ Number of output tokens the model generated as internal reasoning, including
the thinking-block delimiter tokens.
Reflects the raw reasoning the model produced, not the (possibly shorter)
summarized thinking text returned in the response body. Computed by
re-tokenizing the raw reasoning text, so it
caught by the 20-minute sweep ·
the live wire · all receipts