Unverified50% confidenceFactExact time
The minimum cacheable prompt for both 5.5 models is 512 tokens, caching is enabled by marking with cache_control: {"type": "ephemeral"}, and it remains valid for about five minutes.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/17/2026
First Seen
Valid until: 1/15/2027
Sources
Related Entities
Related Claims
UnverifiedThe minimum cacheable length for Haiku 5.5 is 512 tokens, and prompts shorter than this will not error but will not be cached; a cache entry only becomes available after the first response begins; by default the cache lives for 5 minutes and is refreshed for free on each hit.78% similarUnverifiedWhen the context approaches 512K tokens, the prefill-decode disaggregated architecture ensures that users still receive the first response token within 800 milliseconds after reading an entire specification.72% similarUnverifiedThe embedding model's maximum length means that inputs exceeding 512 tokens are silently truncated, so the example sets the chunk budget to 480 tokens.71% similarUnverifiedT5 系列最大支持 512 个 Token 的长序列，而 CLIP 文本编码器最大序列长度通常仅为 77 个 Token70% similarUnverifiedAmong the 600 trajectories in this study, 98.6% of input tokens were provided by the prefix cache.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1054194API
curl https://kongchang.com/api/v1/knowledge/claims/1054194MCP
get_claim(id=1054194)