Head to head
GPT-OSS 20B vs DeepSeek V4 Flash (0731)
On a blended workload of one million input tokens and three million output tokens, GPT-OSS 20B is 1.5× cheaper than DeepSeek V4 Flash (0731).
| GPT-OSS 20B | DeepSeek V4 Flash (0731) | |
|---|---|---|
| Provider | Together AI | Together AI |
| Input / MTok | $0.05 | $0.14 |
| Output / MTok | $0.2 | $0.28 |
| Cache read / MTok | — | — |
| Context window | 128,000 | 1,000,000 |
| Blended 1M in + 3M out | $0.65 | $0.98 |
| Status | ga | ga |