Is DeepSeek-V3.2 (Non-thinking Mode) or DeepSeek V4 Flash cheaper?

DeepSeek V4 Flash is cheaper on output tokens ($0.28 vs $0.42 per 1M).

Which has the larger context window, DeepSeek-V3.2 (Non-thinking Mode) or DeepSeek V4 Flash?

DeepSeek V4 Flash has the larger context window (1M tokens).

DeepSeek-V3.2 (Non-thinking Mode) vs DeepSeek V4 Flash

DeepSeek V4 Flash is cheaper on output tokens, while DeepSeek V4 Flash offers a larger context window. Choose DeepSeek-V3.2 (Non-thinking Mode) or DeepSeek V4 Flash based on the trade-off between cost, context, and the benchmarks that matter for your use case.

Spec	DeepSeek-V3.2 (Non-thinking Mode)	DeepSeek V4 Flash
Provider	ZenMux	ZenMux
Input / 1M tokens	$0.28	$0.14
Output / 1M tokens	$0.42	$0.28
Context window	128K	1M
Parameters	—	284B
Open weights	No	Yes
Released	Dec 2025	Apr 2026

DeepSeek-V3.2 (Non-thinking Mode) details →DeepSeek V4 Flash details →

FAQ

Pricing is indicative — confirm with the provider before production use. Highlighted values indicate the better figure for that row.