#prompt-caching

K3's least-used surfaces are its 1M context, native vision and reasoning-effort dial. Switching effort mid-session forfeits the cache rate that funds it.

Writer's own research reports a harness redesign cut token spend nearly 40% at steady accuracy. Why the orchestration layer, not the model, sets AI cost.

Running Kimi K3 in Kimi Code: the Moderato plan unlocks 256K context, Allegretto the full 1M, and cache discipline — not knobs — controls your real cost.

OpenAI's GPT-5.6 caching overhaul, plus Anthropic and DeepSeek tiers, reshapes agent cost math. How to design cache-first, model-homogeneous agents.

Claude Fable 5 lists at $10/$50 per million tokens and meters after July 7. How prompt caching, the Batch API, and spend caps cut the bill — with worked math.

Uber burned a year's AI budget in four months; Microsoft cut Claude Code. A 4-gate playbook that right-sizes model spend and cuts AI bills 60 to 80 percent.

Prompt caching reuses computed KV tensors so repeated prefixes cost up to 90% less, with no quality loss. The 2026 cross-provider engineering playbook.

Opus 4.8 tops the Artificial Analysis index, but GPT-5.5 still leads Terminal-Bench. An evidence-graded roundup of the first 48 hours of independent evals.

A FinOps playbook for cutting AI inference spend without quality loss: model routing, prompt and KV caching, batching, quantization, and unit-cost tracking.

Tokens, prefill, decode, KV cache, prefix cache, batch tier, reserved capacity, context-window discount — every LLM billing term, defined.

When 1M context pays off — and when it bankrupts you. Token-spend math, prompt-cache strategy, and break-even tables for agentic Claude Opus 4.7 workloads.