Recent candidate changes
DeepSeek API’s Concurrency and Rate Limits
normalized
release_note_update / 87% / no review hint
prompt_cache_miss_tokens: Number of tokens from the input that were not served from the cache ($0.14 per million tokens)
normalized
release_note_update / 96% / no review hint
prompt_cache_hit_tokens：Number of tokens from the input that were served from the cache ($0.014 per million tokens)
normalized
release_note_update / 96% / no review hint
Improve model output performance through Few-shot learning.
normalized
release_note_update / 87% / no review hint
Code analysis and debugging with repeated repository references
normalized
release_note_update / 76% / no review hint
Data analysis with recurring queries on the same documents/files
normalized
release_note_update / 76% / no review hint
Role-play with extensive character settings and multi-turn conversations
normalized
release_note_update / 76% / no review hint
Unused cache entries are automatically cleared, typically within a few hours to days.
normalized
release_note_update / 76% / no review hint
The cache system does not guarantee 100% cache hits.
normalized
release_note_update / 96% / no review hint
The cache system uses 64 tokens as a storage unit; content less than 64 tokens will not be cached.
normalized
release_note_update / 96% / no review hint
The DeepSeek API is designed to handle up to 1 trillion tokens per day, with no limits on concurrency or rate, ensuring
normalized
release_note_update / 96% / no review hint
This is made possible by the MLA architecture in DeepSeek V2, which enhances model performance while significantly reduc
normalized
release_note_update / 96% / no review hint