Measuring Claude 4.7's tokenizer costs

The author used Anthropic's free count_tokens API to compare Claude Opus 4.6 and 4.7 on 7 real samples and 12 synthetic ones; the real-sample weighted total rose from 8,254 to 10,937 input tokens, or 1.325x. Technical docs hit 1.47x, a real CLAUDE.md file hit 1.445x, while Chinese and Japanese stayed near 1.01x. On a 20-prompt IFEval sample, 4.7 improved strict prompt-level pass rate from 85% to 90%; the post cannot isolate tokenizer effects from model weights or post-training.

Why it matters: HKR-H/K/R all land: the post has a sharp cost hook, reproducible token-count data, and clear budget impact for Claude Code users. It stays below p1 because this is a third-party measurement, not an Anthropic release, and the IFEval slice is only 20 items.