Simon Willison 发布 ttok 1.0，默认 tokenizer 切换至 GPT-5/GPT-6
ttok 1.0
Simon Willison 发布 ttok 1.0，将默认 tokenizer 从 GPT-4 切换为 GPT-5/GPT-6。OpenAI 尚未确认 GPT-6 是否沿用 GPT-5 系列的 tokenizer；一项实验中，7 个 GPT 模型在 31 个测试样例上均得到 44,794 个 token，且结果彼此一致。
I released ttok 0.4, ran uv tool upgrade ttok, piped a file into the new version... and realized that it was defaulting to the GPT-4 tokenizer when it should very clearly default to GPT-5/GPT-6 instead!
I figured switching the default was a reasonable excuse to finally ship a 1.0.
OpenAI haven't actually confirmed that GPT-6 uses the same tokenizer as the GPT-5 family yet - there's an angry issue about it - but I found this commit by William Liu which reports on an experiment he ran confirming that the tokenizers are likely the same:
All seven GPT models (5.5, 5.6 Sol/Terra/Luna, 6 Astra/Sol/Luna) report 44,794 tokens and match each other on every one of the 31 fixtures. GPT-6 introduces no input-count change on this corpus.
来源：Simon Willison · simonwillison.net