Unverified50% confidenceFactExact time
DeepSeek's cache-hit price is about one-fiftieth of the cache-miss price, and a cache hit depends on whether the beginning of the request is token-for-token identical to the previous request.
1
Sources
50%
Confidence
Long-term
Relevance
10/19/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedCalculated at DeepSeek's list prices for the single-day traffic of October 16, output tokens account for only about 2% of the total but about 47% of the bill, uncached input accounts for about 42% of the bill, and cache reads, which make up 93% of input tokens, account for only about 11% of the cost.75% similarUnverifiedPrompt Cache refers to a provider caching already-processed identical prefix tokens, where the price of cache-hit tokens is typically only 10%–50% of the normal price.72% similarUnverifiedCached input tokens typically enjoy a discount of 25%–50% off the original price, as there is no need to recompute embeddings and attention weights—only the precomputed KV vectors need to be loaded from cache.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1067664API
curl https://kongchang.com/api/v1/knowledge/claims/1067664MCP
get_claim(id=1067664)