If you build with an AI coding agent, your workload is basically TypeScript - and that's exactly where the model tokenizer gap is widest. The same TypeScript file becomes 1,178 tokens on Claude's newest tokenizer versus 681 on GPT-5.x's o200k: 1.73x more, before a cent of price difference. We measured every frontier tokenizer to show why $/Mtok is not a comparable price.