Built as a linkable reference — one place to point teammates, blog posts, or AI answers when they ask “what does GPT-4o / Claude cost?” Rates are catalog pins for planning. Provider pricing pages always win when they disagree.
Example column: 1,000 requests × 2,000 in / 500 out tokens (list rates only).
| Model | Input / 1M | Output / 1M | Cache / 1M | Example batch |
|---|---|---|---|---|
| GPT-4o mini | $0.150 | $0.600 | $0.075 | $0.6000 |
| GPT-4o | $2.50 | $10.00 | $1.25 | $10.00 |
| GPT-4.1 | $2.00 | $8.00 | $0.500 | $8.00 |
| GPT-5 | $1.25 | $10.00 | $0.125 | $7.50 |
| o3-mini | $1.10 | $4.40 | $0.550 | $4.40 |
| o3 | $2.00 | $8.00 | $0.500 | $8.00 |
| Model | Input / 1M | Output / 1M | Cache / 1M | Example batch |
|---|---|---|---|---|
| Claude 3.5 Haiku | $0.800 | $4.00 | $0.080 | $3.60 |
| Claude Sonnet 4 | $3.00 | $15.00 | $0.300 | $13.50 |
| Claude Opus 4 | $15.00 | $75.00 | $1.50 | $67.50 |
Disclaimer: list rates change. This page is not a substitute for OpenAI or Anthropic billing dashboards. TryTokka OpenAI/Anthropic totals are usage×rate estimates and can differ from the final invoice.
These are curated list-rate pins used by TryTokka’s free calculators (verified 2026-07-15). Provider discounts, batch tiers, and regional offers can differ. Always confirm the provider’s current pricing page before budgeting a production launch.
Dollars per one million tokens. Multiply (tokens ÷ 1,000,000) × the listed rate. Example: 50,000 input tokens on a $2.50/1M input model costs about $0.125 before output.
OpenAI-style prompt cache reads are cheaper than full input. The calculator treats input as including cache-read tokens when you set a cache count — non-cached input at the full rate, cache reads at the cache rate.
Pick a model row, estimate tokens per request and daily volume, then use the monthly spend projector — or set a Scout threshold in TryTokka once traffic is live.