Cacheable Prefix Prompt Layout

Data artifactModel ServingModelsarc:CacheablePrefixPromptLayout

A prompt-structure convention that places static, frequently reused content (system instructions, tool definitions, stable user profile) first and variable request-specific content last so provider prefix caches can match it.

Responsibility. Orders prompt segments by cacheability to maximize prefix-cache hits.

Also known as: Cache-friendly prompt structure, Static-first prompt ordering

Direct neighbourhood (hover for relationship types)

Relationships

configures structural

Prompt Context Builder Ch8.3

Design guidance

MUST place static, frequently reused content at the beginning of the prompt, followed by variable, request-specific content.
SHOULD NOT include content that changes more often than it is reused (e.g., daily market data) in the cacheable prefix.

Quantitative guidance

As stated by the sources; verify before use.

Financial advisor: cacheable prefix of 5,800 tokens (800 system prompt + 5,000 client profile) out of 9,000 input tokens; market data (3,000) and query (200) remain variable (Ch8.3).
Regulatory context of 12,000 tokens (SEC 3,500, FINRA 4,200, DOL 2,800, state 1,500) repeated in 20,000 of 50,000 monthly requests = 240M input tokens ($600/month) of redundant transmission (Ch8.3).

Classification

Patterns: Prompt cachingPrefix matching
Quality attributes: Performance efficiency (ISO/IEC 25010)
Risks mitigated: Cache misses caused by variable content inside the prompt prefixRedundant retransmission of static context

Sources

Ch8.3: T. Nguyen, "Token Economics and Architecture," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 8.3. ISBN: 9798244538229.