Prompt caching correctly — the variable input goes LAST — step 8 of 9

Write estimate_cost(stable_tokens, variable_tokens, calls, cached, input_rate) that:

• Returns the total cost in dollars (a float, rounded to 2 places).
• When cached=False: input_rate * (stable_tokens + variable_tokens) * calls / 1_000_000.
• When cached=True:
  • Cache write (one-time): input_rate * 1.25 * stable_tokens / 1_000_000
  • Cache reads (calls - 1 of them): input_rate * 0.1 * stable_tokens * (calls - 1) / 1_000_000
  • Variable tokens (full price every call): input_rate * variable_tokens * calls / 1_000_000
  • Sum them.

Two cases run for you. Expected output:

uncached: 2.46
cached: 0.33
savings: 2.13