OpenAI
Per-million-token prices as published on the vendor’s own pricing page (USD). Cache-hit rates and peak/off-peak tiers are vendor-specific and usually absent from routed-price listings.
| Model | Input /M | Output /M | Cache hit | Peak | Context | Source |
|---|---|---|---|---|---|---|
| GPT-6.1 Sol mainstream | $2 | $10 | — | — | 1.05M Tokens | official ↗ |
| GPT-6 Astra intlFlagship | $10 | $50 | — | — | 1.1M Tokens | official ↗ |
| GPT-5.6 Luna intlFlagship | $0.2 | $1.20 | — | — | 1.1M Tokens | official ↗ |
| GPT-5.6 Sol mainstream | $2 | $10 | — | — | 1.05M Tokens | official ↗ |
| GPT-6 Luna mainstream | $0.1 | $0.5 | — | — | 1.05M Tokens | official ↗ |
| GPT-5.6 Terra mainstream | $2 | $12 | — | — | 1.05M Tokens | official ↗ |
| GPT-5 Nano mainstream | $0.05 | $0.4 | — | — | 400,000 Tokens | official ↗ |
Peak price applies only to vendors that bill by time of day (e.g. DeepSeek off-peak discount). Cache-hit price applies when the prompt prefix is cached.