Token prices vary by over 100× across models. A quick reference (approximate, USD per 1M tokens — always confirm on the provider's page):
| Model | Input | Output |
|---|---|---|
| gpt-4o | $2.5/M | $10/M |
| gpt-4o-mini | $0.15/M | $0.6/M |
| gpt-4.1 | $2/M | $8/M |
| gpt-4.1-nano | $0.1/M | $0.4/M |
| o3 | $10/M | $40/M |
| claude-opus-4 | $15/M | $75/M |
| claude-sonnet-4 | $3/M | $15/M |
| claude-haiku-4 | $1/M | $5/M |
| gemini-2.5-flash | $0.15/M | $0.6/M |
The lesson: match the model to the task. The token optimizer compares your exact prompt across all of these and names the cheapest.