Head to head
Qwen2.5 7B Instruct Turbo vs DeepSeek V4 Flash
On a blended workload of one million input tokens and three million output tokens, DeepSeek V4 Flash is 1.2× cheaper than Qwen2.5 7B Instruct Turbo.
| Qwen2.5 7B Instruct Turbo | DeepSeek V4 Flash | |
|---|---|---|
| Provider | Together AI | DeepSeek |
| Input / MTok | $0.3 | $0.14 |
| Output / MTok | $0.3 | $0.28 |
| Cache read / MTok | — | $0.0028 |
| Context window | 32,768 | 1,000,000 |
| Blended 1M in + 3M out | $1.20 | $0.98 |
| Status | ga | ga |