Take audit: sx_lCBkUwV_265xulu57o
2026-08-22T09:19:10.751Z — Mika
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 · 647 ms · verdict: unjudged
(empty)
mini_auto_bare_deg gpt-4o-mini-transcribe · degraded · auto · bare
— MER accuracy pending Ground Truth · 587 ms · verdict: unjudged
(empty)
mini_auto_bare_clean gpt-4o-mini-transcribe · clean · auto · bare
— MER accuracy pending Ground Truth · 606 ms · verdict: unjudged
(empty)
mini_auto_prompt_deg gpt-4o-mini-transcribe · degraded · auto · mixed prompt
— MER accuracy pending Ground Truth · 1006 ms · verdict: unjudged
(empty)
mini_auto_prompt_clean gpt-4o-mini-transcribe · clean · auto · mixed prompt
— MER accuracy pending Ground Truth · 791 ms · verdict: unjudged
以下是繁體中文與英文混合語音，請原樣轉錄，保留英文單字與繁體中文，不要翻譯。
mini_auto_prompt_64k gpt-4o-mini-transcribe · 64k · auto · mixed prompt
— MER accuracy pending Ground Truth · 874 ms · verdict: unjudged
(empty)
mini_zh_prompt_deg gpt-4o-mini-transcribe · degraded · zh · mixed prompt
— MER accuracy pending Ground Truth · 938 ms · verdict: unjudged
(empty)
mini_zh_prompt_clean gpt-4o-mini-transcribe · clean · zh · mixed prompt
— MER accuracy pending Ground Truth · 1083 ms · verdict: unjudged
以下是繁體中文與英文混合語音，請原樣轉錄，保留英文單字與繁體中文，不要翻譯。
gpt_tw_en_bare_deg gpt-transcribe · degraded · zh-tw + en · bare
— MER accuracy pending Ground Truth · 477 ms · verdict: unjudged
(empty)
gpt_tw_en_prompt_deg gpt-transcribe · degraded · zh-tw + en · mixed prompt
— MER accuracy pending Ground Truth · 745 ms · verdict: unjudged
(empty)
gpt_tw_en_prompt_clean gpt-transcribe · clean · zh-tw + en · mixed prompt
— MER accuracy pending Ground Truth · 570 ms · verdict: unjudged
(empty)
gpt_tw_en_prompt_64k gpt-transcribe · 64k · zh-tw + en · mixed prompt
— MER accuracy pending Ground Truth · 692 ms · verdict: unjudged
(empty)
gpt_tw_only_prompt_deg gpt-transcribe · degraded · zh-tw · mixed prompt
— MER accuracy pending Ground Truth · 622 ms · verdict: unjudged
(empty)
Winning configuration
Record the variant and rationale to hand to qa-rails after reviewing this Take.