codex truncated tool calls

codex truncated tool calls

apparently it’s because the tokenizer isn’t open source, isn’t compatible with the one that tiktoken ships

the API call to measure length is too slow, and they don’t want to walk the line too close, so they just truncate

from what i’ve seen, gpt5 in codex will aggressively self-limit, e.g. using `sed -n 10,20p` to read small chunks instead of the full file.

so not sure how bad of a problem it is in practice