Take audit: BdZqQDq_rbAtfohOj2V1t
2026-08-22T09:34:33.754Z — 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 · 622 ms · verdict: unjudged
我有這一本書,Can you see?
mini_auto_bare_deg gpt-4o-mini-transcribe · degraded · auto · bare
— MER accuracy pending Ground Truth · 485 ms · verdict: unjudged
我有這一本書,can you see?
mini_auto_bare_clean gpt-4o-mini-transcribe · clean · auto · bare
— MER accuracy pending Ground Truth · 653 ms · verdict: unjudged
我有注意一本書,Can you see?
mini_auto_prompt_deg gpt-4o-mini-transcribe · degraded · auto · mixed prompt
— MER accuracy pending Ground Truth · 590 ms · verdict: unjudged
我有three本書,can you see?
mini_auto_prompt_clean gpt-4o-mini-transcribe · clean · auto · mixed prompt
— MER accuracy pending Ground Truth · 687 ms · verdict: unjudged
One more tree本書Can you see?
mini_auto_prompt_64k gpt-4o-mini-transcribe · 64k · auto · mixed prompt
— MER accuracy pending Ground Truth · 853 ms · verdict: unjudged
我有三本書,can you see?
mini_zh_prompt_deg gpt-4o-mini-transcribe · degraded · zh · mixed prompt
— MER accuracy pending Ground Truth · 668 ms · verdict: unjudged
我有three本書,can you see?
mini_zh_prompt_clean gpt-4o-mini-transcribe · clean · zh · mixed prompt
— MER accuracy pending Ground Truth · 717 ms · verdict: unjudged
我有這一本書,can you see?
gpt_tw_en_bare_deg gpt-transcribe · degraded · zh-tw + en · bare
— MER accuracy pending Ground Truth · 762 ms · verdict: unjudged
我有三本書，can you see?
gpt_tw_en_prompt_deg gpt-transcribe · degraded · zh-tw + en · mixed prompt
— MER accuracy pending Ground Truth · 747 ms · verdict: unjudged
我有三本書，can you see?
gpt_tw_en_prompt_clean gpt-transcribe · clean · zh-tw + en · mixed prompt
— MER accuracy pending Ground Truth · 820 ms · verdict: unjudged
我有三本書，can you see?
gpt_tw_en_prompt_64k gpt-transcribe · 64k · zh-tw + en · mixed prompt
— MER accuracy pending Ground Truth · 626 ms · verdict: unjudged
我有三本書，can you see?
gpt_tw_only_prompt_deg gpt-transcribe · degraded · zh-tw · mixed prompt
— MER accuracy pending Ground Truth · 563 ms · verdict: unjudged
我有三本書，can you see?
Winning configuration
Record the variant and rationale to hand to qa-rails after reviewing this Take.