Take audit: tW-C9hmxJKLjrNJfl07Sk

2026-08-22T09:19:16.162Z — ray

0/13 judged · keyboard: Space play/pause, ↑/↓ focus, P pass, T Translation failure, F fail

Do a blind check (transcripts hidden)

Ground Truth

What the Tester actually said. Never pre-filled from an Engine — type what you hear.


Transcripts

mini_prod_baseline gpt-4o-mini-transcribe · degraded · en · bare — MER accuracy pending Ground Truth · 554 ms · verdict: unjudged

Thank you,老師謝謝你

mini_auto_bare_deg gpt-4o-mini-transcribe · degraded · auto · bare — MER accuracy pending Ground Truth · 646 ms · verdict: unjudged

Thank you,老師謝謝你

mini_auto_bare_clean gpt-4o-mini-transcribe · clean · auto · bare — MER accuracy pending Ground Truth · 512 ms · verdict: unjudged

Thank you,老師謝謝你

mini_auto_prompt_deg gpt-4o-mini-transcribe · degraded · auto · mixed prompt — MER accuracy pending Ground Truth · 942 ms · verdict: unjudged

Thank you, 老師謝謝你

mini_auto_prompt_clean gpt-4o-mini-transcribe · clean · auto · mixed prompt — MER accuracy pending Ground Truth · 998 ms · verdict: unjudged

Thank you, 老師謝謝你

mini_auto_prompt_64k gpt-4o-mini-transcribe · 64k · auto · mixed prompt — MER accuracy pending Ground Truth · 811 ms · verdict: unjudged

Thank you, 老師謝謝你

mini_zh_prompt_deg gpt-4o-mini-transcribe · degraded · zh · mixed prompt — MER accuracy pending Ground Truth · 970 ms · verdict: unjudged

Thank you, 老師謝謝你

mini_zh_prompt_clean gpt-4o-mini-transcribe · clean · zh · mixed prompt — MER accuracy pending Ground Truth · 713 ms · verdict: unjudged

Thank you,老師謝謝你

gpt_tw_en_bare_deg gpt-transcribe · degraded · zh-tw + en · bare — MER accuracy pending Ground Truth · 803 ms · verdict: unjudged

Thank you，老師謝謝你。

gpt_tw_en_prompt_deg gpt-transcribe · degraded · zh-tw + en · mixed prompt — MER accuracy pending Ground Truth · 611 ms · verdict: unjudged

Thank you，老師謝謝你。

gpt_tw_en_prompt_clean gpt-transcribe · clean · zh-tw + en · mixed prompt — MER accuracy pending Ground Truth · 615 ms · verdict: unjudged

Thank you，老師謝謝你。

gpt_tw_en_prompt_64k gpt-transcribe · 64k · zh-tw + en · mixed prompt — MER accuracy pending Ground Truth · 756 ms · verdict: unjudged

Thank you，老師謝謝你。

gpt_tw_only_prompt_deg gpt-transcribe · degraded · zh-tw · mixed prompt — MER accuracy pending Ground Truth · 798 ms · verdict: unjudged

Thank you，老師謝謝你。

Winning configuration

Record the variant and rationale to hand to qa-rails after reviewing this Take.