I Benchmarked 6 LLMs to Automate My Job Board for $0.35/Month

Update (April 11, 2026): I re-ran benchmarks with March 2026 models. GPT-5.4 Nano (\(0.20/\)1.25 per M tokens) matched GPT-5 Mini on classification (96% F1) and beat Claude Haiku on extraction (100% s