WOLF / CODEX OBSERVATORY
LOCAL HISTORY · API SIMULATION
YOUR COMPUTE, IN PERSPECTIVE

What would this
cost on the API?

Actual Codex token usage. Estimated API-equivalent spend.
No inference calls. No chat content.

API-EQUIVALENT COST—Estimated at current Standard rates
TOTAL TOKENS——
CACHE HIT RATE—Cached tokens / all input tokens
OUTPUT TOKENS—Includes reasoning; counted once

Usage over time

API-equivalent USD

Model breakdown

ModelTokensAPI equivalent

Token composition

Cached input is billed at the discounted rate, not the full input price.

Pricing & methodology

This is a hypothetical API token bill, not money charged to your account or subscription savings. Rates are current as of , applied to all recorded history. Subscription fees, tax, regional uplifts, paid tools, Fast, Batch and Flex are excluded.

Uncached input = input − cached input − cache writes. Reasoning is already included in output. Long-context multipliers apply above 272,000 input tokens: 2× input/cache and 1.5× output. GPT-5.5 applies that multiplier to the affected recorded session.

Response records are preferred; older logs use last-request token counters when cumulative totals advance. Repeated records are deduplicated across files. Data is aggregated into one-minute buckets; rolling filters use those timestamps. Missing local history is not treated as proof of zero account usage.