Qwen3 and GPT Both Fail to Count the e's in 'seventeen'

A viral comparison screenshot shows both Qwen3 and GPT incorrectly answering how many e's are in 'seventeen', with the local Qwen3 model replying '4' after misspelling the word letter by letter, highlighting LLMs' persistent weakness at character-level tasks.

2026-10-31 ~ 2026-10-31 · 2 related posts

• Qwen and ChatGPT both flub counting the e's in 'seventeen' — Express_Quail_1493 · 2026-10-31
• Counting the E's in 'Seventeen': Qwen3 and GPT Disagree, Both Wrong — Express_Quail_1493 · 2026-10-31