Blog

Notes, comparisons, and tactics for shipping cost-efficient AI products.

Published March 4, 20254 min read

How to cut your AI API costs by 80%

Eight concrete techniques that have repeatedly cut LLM API bills by 50–90% in production, ranked by effort-to-payoff. No vague advice.

Published February 20, 20252 min read

GPT-4o vs Claude 3.5 Sonnet: price and performance

A head-to-head comparison of the two most-used mid-flagship models in production: pricing, context windows, strengths, weaknesses, and which one to pick for which workload.

Published February 12, 20253 min read

Cheapest LLM APIs in 2025: a complete comparison

A practical, no-fluff comparison of the cheapest production-grade LLM APIs in 2025, with input/output prices, context windows, and which ones to actually pick for your workload.