API pricing comparison
Claude vs GPT API pricing
As of 2026-06-22, Anthropic opus-4-8 costs $5 per 1M input and $25 per 1M output tokens, while OpenAI gpt-5.5 costs $5 input and $30 output. Each price is dated and first-party-sourced.
| Model | Provider | Input /1M | Output /1M |
|---|---|---|---|
| opus-4-8 | Anthropic | $5 | $25 |
| sonnet-4-6 | Anthropic | $3 | $15 |
| haiku-4-5 | Anthropic | $1 | $5 |
| gpt-5.5 | OpenAI | $5 | $30 |
| gpt-4o | OpenAI | $2.50 | $10 |
| Series | Date | Value |
|---|---|---|
| opus-4-8 | 2026-05-28 | $5 |
| gpt-5.5 | 2026-04-24 | $5 |
| sonnet-4-6 | 2026-02-17 | $3 |
| gpt-4o | 2024-05-13 | $5 |
| gpt-4o | 2024-08-06 | $2.50 |
How to read this
These are standard published list prices per 1M tokens, not promotional, batch, or cached rates, and not the same as what you actually pay once caching and short prompts are taken into account. Output tokens cost several times input, so the model mix and how chatty each model is matter as much as the sticker price. To see the price you were actually charged on a given day, priced at the rate in effect then, use Goei or the claude-code-cost plugin.
More from the index
- All per-model pricing pages
- The cheapest LLM API prices per token
- The interactive chart
- The dataset, JSON endpoints, and how to cite it
Open data, CC BY 4.0. Prices validated as of 2026-06-22.