Blog
Practical writing on LLM costs, AI budgets, and keeping API spend under control.
October 11, 2026 · 3 min read
Model Routing Can Cut LLM Costs 85%. The Real Evidence
RouteLLM's benchmark found model routing cuts LLM costs up to 85% while keeping 95% of GPT-4's quality. Here is what vendor case studies get wrong.
Read the postSeptember 19, 2026 · 3 min read
Why Did My LLM Bill Spike? Five Runaways, Priced by the Hour
Four agents looped for 11 days and ran up $47,000. The five ways an LLM bill spikes, what each costs per hour, and how fast you would catch it.
Read the postSeptember 13, 2026 · 3 min read
Amazon Missed a $1.8 Million AI Overrun for Five Months. Here's the AI Spend Visibility Gap Behind It.
Amazon ran $1.8M over budget on one Claude task and missed it for five months. The AI spend visibility gap behind it, and what to check this week.
Read the postAugust 15, 2026 · 3 min read
79% of Companies Had an AI Budget Overrun This Year — Here's Why Falling Prices Won't Save You
79% of enterprises had an AI budget overrun this year. Here's why falling model prices won't fix it, and what small businesses can do this week.
Read the postJuly 25, 2026 · 4 min read
Four Frontier Models in 17 Days: Why the AI Price War Won't Cut Your LLM Costs
Grok 4.5, GPT-5.6, Gemini 3.6 Flash and Claude Opus 5 all shipped in one July — with prices falling. So why did 79% of finance leaders report AI budget overruns? Because LLM costs and LLM prices are different numbers.
Read the post