Fine-tuning cost

Two bills, one project: a one-time training run and a monthly serving tab. The trap is assuming the training bill is the big one — for most teams, serving dwarfs it within a quarter.

Training examples

Average tokens per example (prompt + completion, both count)

Epochs

Training price $/1M tokens check your vendor's current fine-tuning rate

Serving: input tokens / month (millions)

Serving: output tokens / month (millions)

Serving model (your fine-tuned model bills at base-model rates)

The formula

train_tok = examples × tokens_per_example × epochs training = train_tok/1M × training_price (one-time) serving = sIn × P_in + sOut × P_out (monthly) break_even = training / monthly_saving (if tuning cuts your prompt size)

Training-rate default is an editable placeholder in the typical hosted range ($2–4/1M tokens as of ) — vendors change it, so paste yours. Serving bills at base-model per-token rates via OpenRouter's daily pull. Fine-tuning pays off only when it lets you shrink prompts (bake instructions in) or downgrade the base model — model the saving on the chatbot calculator for comparison.