> From: Johan Ekenberg <johan.ekenberg@gmail.com>
> Date: Tue, 7 Jan 2025 21:08:09 +0100
>
> - Start emacs: 'emacs -Q'
> - Paste this function into the scratch buffer and evaluate it:
>   (defun test-org-insertion-fontification ()
>     (interactive)
>     (let* ((chunks
>             '("#+begin_src elisp\n"
>               "(+ 1 2)"
>               "\n#+end_src\n")))
>       (insert "* NOTE: You need to run this with org-mode enabled and point at end of buffer\n\n")
>       (insert "** This code block is fontified correctly\n")
>       (insert (car chunks))
>       (dolist (chunk (cdr chunks))
>         (insert chunk))
>
>       (insert "\n** But not this code block\n")
>       (insert (car chunks))
>       (dolist (chunk (cdr chunks))
>         (sit-for (/ 10 1000.0)) ; simulate chunks received over network and streamed into buffer
>         ;; The following were also tested one by one instead of the sit-for line above:
>         ;; (redisplay) ; this triggers the bug
>         ;; (force-mode-line-update) ; this does not trigger the bug
>         ;; (read-event nil nil 0) ; this triggers the bug
>         (insert chunk))
>       ))
> - Open an existing or new file/buffer with org-mode enabled (C-x C-f foo.org)
> - Place point at end of buffer
> - M-x test-org-insertion-fontification
> - Optional: replace the line (sit-for...) with any of the lines below it.
>
> - Additional information: The problem was discovered when receiving
> streamed data responses via the package gptel which is an Emacs LLM
> Client. Here is the reported issue:
> https://github.com/karthink/gptel/issues/550
>
> - Also replicated on Emacs 29.3

Isn't this an Org problem?  If so, could you please report this to the
Org bug mailing list?