SpeechEditBench - Paralinguistic: leaderboard
SpeechEditBench (2026) tests speech language models on editing a recording from a natural-language instruction in English and Mandarin, scored by automatic checks and an audio judge. This board covers adding or removing breaths, laughs, coughs and sighs. The score is joint success in percent: the edit must succeed and the spoken words must be preserved; higher is better.
Metric: Joint success (%): the requested edit is applied (task-specific automatic check) and the linguistic content is preserved (Whisper large-v3 or Paraformer transcript within 10% WER or CER of the expected text); source speech plus a natural-language instruction, English and Mandarin samples, each model through its released editing or conversation interface; paralinguistic editing (add or remove a breath, laugh, cough or sigh, judged by a gemini-2.5-pro audio judge), 400 samples; higher is better. Source: arxiv.org. Saturation forecast: Around September 2027. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen3 Omni 30B A3B Instruct | 14.5 | #904 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=speecheditbench-paralinguistic · How It Works · Data refreshed daily, snapshot 2026-11-13.