Head to head

DeepSeek V4 Flash vs Qwen2.5 7B Instruct Turbo

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.

DeepSeek V4 FlashQwen2.5 7B Instruct Turbo
Provider DeepSeek Together AI
Input / MTok $0.14 $0.3
Output / MTok $0.28 $0.3
Cache read / MTok $0.0028
Context window 1,000,000 32,768
Blended 1M in + 3M out $0.98 $1.20
Status ga ga

DeepSeek V4 Flash detail   Qwen2.5 7B Instruct Turbo detail