LLM Fine-Tuning Cost Estimator
Plan one-off token-based training and monthly inference or hosting costs with editable rates. Keep training and serving budgets separate.
Your estimate
- One-off training
- $1.50
- Monthly inference
- $1.60
- Monthly total with hosting
- $1.60
- Training + first month
- $3.10
Illustrative custom rates, not a quote or a model availability claim. No model is selected. Replace each rate with your provider's current fine-tuned rates; base-model prices may differ.
How it works
Training cost = dataset tokens x epochs x training rate / 1 million. Monthly serving = input token cost + output token cost + your fixed hosting budget. The first-month estimate adds training once. Rates are illustrative custom inputs, not provider presets.
Know the limits
This is a token-priced scenario estimator, not a GPU-hour training model. It excludes evaluation, failed runs, data preparation, storage, tax and transfer. Provider eligibility matters: OpenAI documentation checked November 14, 2026 says fine-tuning is closed to new users and being wound down. Confirm availability and model shutdown dates before planning a job.
Reference sources
Runs in your browser. Inputs are not uploaded or saved. Explore all free tools or read the technical notes.