Head to head
DeepSeek V4 Flash (0731) vs GPT-5 nano
On a blended workload of one million input tokens and three million output tokens, DeepSeek V4 Flash (0731) is 1.3× cheaper than GPT-5 nano.
| DeepSeek V4 Flash (0731) | GPT-5 nano | |
|---|---|---|
| Provider | Together AI | OpenAI |
| Input / MTok | $0.14 | $0.05 |
| Output / MTok | $0.28 | $0.4 |
| Cache read / MTok | — | $0.005 |
| Context window | 1,000,000 | — |
| Blended 1M in + 3M out | $0.98 | $1.25 |
| Status | ga | ga |