O04: Automatic Prefix Caching

• Publisher: vLLM
• URL: Primary source
• Published: not stated
• Updated/version date: 2026-10-13
• Accessed: 2026-10-22
• Type: official runtime documentation, rolling latest
• Verification: opened source, rather than relying on a search excerpt

Supported finding

Shared-prefix KV reuse avoids repeated prefill. It does not reduce the decoding work needed to generate new tokens.

Limits and transfer

Requires matching reusable prefixes. No financial saving is guaranteed for a fixed-capacity deployment.

Use in the session

Mechanism and decision conditions. Retain the qualifications when translating the finding into slide copy. No source establishes Knauf’s savings.