Appendix A showed token-class inputs and a final cost_usd/aic result, but omitted the §3.5 adjustment needed to reproduce that result. As written, a straightforward per-class calculation produced a higher total than the published example.

• Problem

  • The worked example silently depended on subtracting cache_read_tokens from the reported input_tokens before applying the input price.
  • That made the appendix non-reproducible as a conformance check.

• Changes

  • Updated Appendix A to state that the example assumes cache-read tokens are bundled into the reported input total, so §3.5 applies.
  • Renamed the input line to Input (raw) to distinguish the provider-reported total from the billable net input.
  • Added an explicit deduction step showing 1000 - 400 = 600 net input tokens.
  • Added a per-class cost breakdown so the published cost_usd = 0.0054825 result can be verified directly.

• Worked example

  Step 1 — Apply §3.5: net input = 1000 − 400 = 600 tokens
  
  input:       600 × 0.000003    = 0.0018000
  output:      200 × 0.000015    = 0.0030000
  cache_read:  400 × 0.0000003   = 0.0001200
  cache_write:  50 × 0.00000375  = 0.0001875
  reasoning:    25 × 0.000015    = 0.0003750