At current regular rates, DeepSeek V4 Flash costs $0.0028 per 1M cache-hit input tokens, $0.14 per 1M cache-miss input tokens, and $0.28 per 1M output tokens; total cost is the sum of those three usage buckets.
Divide each token count by one million, multiply by its published rate, then add the results. Cache creation or uncached prompt tokens belong to the cache-miss bucket; only reported cache-read tokens receive the hit rate.
cost = (cache_hit_tokens / 1M × $0.0028)
+ (cache_miss_tokens / 1M × $0.14)
+ (output_tokens / 1M × $0.28)For 8M input tokens at a 75% cache-hit ratio and 2M output tokens: 6M hits cost $0.0168, 2M misses cost $0.28, and output costs $0.56. Total: $0.8568.
DeepSeek has announced a future peak/off-peak policy that may double prices during peak windows, but its pricing page says the effective date is subject to a later announcement. This page therefore uses the currently effective regular rates.
Current regular pricing is $0.0028 per 1M cache-hit input tokens, $0.14 per 1M cache-miss input tokens, and $0.28 per 1M output tokens.
Only input tokens that the provider reports as cache reads receive the $0.0028 hit rate. Uncached and cache-creation input use the miss rate.
The official page announces a future peak/off-peak policy but says its effective date will be announced later, so this guide uses current regular rates.
