TaskPriced

TaskPriced is an independent AI cost comparison platform focused on real per-task expenses for Claude and GPT models. It is positioned for developers, engineering leaders, product managers, AI operations teams, and procurement or finance staff who need to plan budgets, choose models, and understand how API pricing behaves under production workloads. The core features include model-by-model cost comparisons, a detailed cost calculator covering batch, flex, fast, and ultrafast tiers, cache reads and writes, long-context thresholds, and monthly volume; a local CSV usage audit tool; a methodology page; and model update tracking. Content is source-linked, dated, and benchmark-aware, comparing only scores measured in the same documented run. It highlights long-context cliffs, reasoning-token effects, and other hidden drivers that make per-token list prices misleading. User experience is direct: browse comparison cards by model or workload, jump into the calculator, or audit an existing usage export. Technical features include dark and light themes, official vendor pricing verification, tier-aware arithmetic, unsupported combination detection, and evidence-graded recommendations labeled Measured, Inferred, or Insufficient evidence.