Models Pricing
Pay per token. No subscriptions, no minimums - you are billed only for what your requests actually use.
| Model | Input price | Cached input | Cache write | Output price | Context | Estimated cost |
|---|---|---|---|---|---|---|
| DeepSeek V3.2 | $0.186 / 1M tokens | - | - | $0.280 / 1M tokens | 128K | $0.026 / request |
| DeepSeek V4 Flash | $0.081 / 1M tokens | $0.007 / 1M tokens | - | $0.162 / 1M tokens | 1M | $0.082 / request |
| DeepSeek V4 Flash 0731 | $0.200 / 1M tokens | $0.007 / 1M tokens | - | $0.500 / 1M tokens | 1M | $0.202 / request |
| DeepSeek V4 Pro | $0.392 / 1M tokens | $0.128 / 1M tokens | - | $0.783 / 1M tokens | 1M | $0.395 / request |
| DeepSeek V4.1 FlashPromo | $0.130 / 1M tokens | $0.003 / 1M tokens | - | $0.500 / 1M tokens | 1M | $0.132 / request |
| GPT-5.6 Luna | $0.180 / 1M tokens | $0.020 / 1M tokens | - | $1.100 / 1M tokens | 1M | $0.185 / request |
| GPT-5.6 Sol | $2.400 / 1M tokens | $0.200 / 1M tokens | - | $14.000 / 1M tokens | 1M | $2.457 / request |
| GPT-5.6 Terra | $1.800 / 1M tokens | $0.200 / 1M tokens | - | $10.000 / 1M tokens | 1M | $1.841 / request |
| GPT-6 AstraPromo | $8.000 / 1M tokens | $1.000 / 1M tokens | - | $30.000 / 1M tokens | 1M | $8.123 / request |
| GPT-6 Luna | $0.100 / 1M tokens | $0.010 / 1M tokens | - | $0.400 / 1M tokens | 1M | $0.102 / request |
| GPT-6 SolPromo | $1.800 / 1M tokens | $0.200 / 1M tokens | - | $8.000 / 1M tokens | 1M | $1.833 / request |
| GPT-6.1 Sol | $2.000 / 1M tokens | $0.100 / 1M tokens | - | $10.000 / 1M tokens | 1M | $2.041 / request |
| Kimi K2.6 | $0.581 / 1M tokens | $0.128 / 1M tokens | - | $2.448 / 1M tokens | 262K | $0.162 / request |
| Kimi K2.7 Code | $0.639 / 1M tokens | $0.152 / 1M tokens | - | $3.150 / 1M tokens | 262K | $0.180 / request |
| Kimi K3 | $2.500 / 1M tokens | $0.300 / 1M tokens | - | $14.000 / 1M tokens | 1M | $2.679 / request |
| GLM 5.3 | $1.170 / 1M tokens | $0.245 / 1M tokens | - | $3.960 / 1M tokens | 1M | $1.243 / request |
| Claude Fable 5.1 | $10.000 / 1M tokens | $0.150 / 1M tokens | $7.500 / 1M tokens | $50.000 / 1M tokens | 1M | $10.205 / request |
| Claude Opus 5.5 | $4.000 / 1M tokens | $0.120 / 1M tokens | $3.000 / 1M tokens | $20.000 / 1M tokens | 1M | $4.082 / request |
| Claude Sonnet 5.5 | $2.000 / 1M tokens | $0.120 / 1M tokens | $1.500 / 1M tokens | $10.000 / 1M tokens | 1M | $2.041 / request |
| Claude Haiku 5.5 | $0.100 / 1M tokens | $0.006 / 1M tokens | $0.075 / 1M tokens | $0.500 / 1M tokens | 1M | $0.102 / request |
Cached input is the discounted rate applied to prompt-cache hits; models without caching show a dash. Cache write is the rate for tokens written into the prompt cache, where a model bills that separately; all other models show a dash. Estimated cost is calculated per request from the table: the full context window as input plus a typical 4K-token response, at the listed input and output prices. Models marked Promo are at a temporary promotional rate.
Image generation
| Model | Price |
|---|---|
| Flux 1 Schnell | $0.002 / image |
| Flux Pro 1.1 | $0.040 / image |
GPT Image models price by quality and output size (USD per image).
GPT Image 2
| Quality | 1024x1024 | 1024x1536 | 1536x1024 |
|---|---|---|---|
| low | $0.005 | $0.004 | $0.004 |
| medium | $0.043 | $0.031 | $0.031 |
| high | $0.151 | $0.125 | $0.125 |
GPT Image 1.5
| Quality | 1024x1024 | 1024x1536 | 1536x1024 |
|---|---|---|---|
| low | $0.008 | $0.009 | $0.009 |
| medium | $0.024 | $0.040 | $0.040 |
| high | $0.113 | $0.100 | $0.100 |