Head to head
DeepSeek V4 Flash (0731) vs Ministral 3 14B
On a blended workload of one million input tokens and three million output tokens, Ministral 3 14B is 1.2× cheaper than DeepSeek V4 Flash (0731).
| DeepSeek V4 Flash (0731) | Ministral 3 14B | |
|---|---|---|
| Provider | Together AI | Mistral |
| Input / MTok | $0.14 | $0.2 |
| Output / MTok | $0.28 | $0.2 |
| Cache read / MTok | — | — |
| Context window | 1,000,000 | — |
| Blended 1M in + 3M out | $0.98 | $0.80 |
| Status | ga | ga |