SpeechEditBench
Models edit English and Chinese speech from written instructions. Joint success requires the requested change and preservation of the attributes checked for that task.
The release contains 4,700 requests. The dots.tts.edit paper tested 3,680 supported requests. The AuK paper reports joint success for five edit types.
- Made by
- City University of Hong Kong and collaborators
- Released
- April 2026
- Kind
- Speech recognition and translation
- Level
- University
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Capability
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Audio and speech
- Last model added
- August 2025
- Scores refreshed
- November 21st, 2026
gpt-realtime leads with 96.7% gpt-realtime leads with 96.7%
Best score per entry from the benchmark's paper, read on November 21st, 2026. A paper’s table doesn’t add newer models. The paper reports no overall score, so this is the content-editing joint-success column, which it presents first. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.
- 1 gpt-realtime 96.7%
- 2 Gemini-Live 93.2%
- 3 Ming-UniAudio 76.5%
- 4 Qwen3-Omni 72%
- 5 Mimo-Audio-Instruction 64.2%
- 6 Kimi-Audio-7B-Instruct 34.7%
- 7 Mimo-Audio-Base 31.7%
- 8 Step-Audio-EditX 16.5%
Source: The SpeechEditBench paper
Source: The SpeechEditBench paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | gpt-realtime | 96.7% | Aug 2025 | Official leaderboard | |
| 2 | Gemini-Live | 93.2% | Official leaderboard | ||
| 3 | Ming-UniAudio | 76.5% | Official leaderboard | ||
| 4 | Qwen3-Omni | 72% | Official leaderboard | ||
| 5 | Mimo-Audio-Instruction | 64.2% | Official leaderboard | ||
| 6 | Kimi-Audio-7B-Instruct | 34.7% | Apr 2025 | Official leaderboard | |
| 7 | Mimo-Audio-Base | 31.7% | Official leaderboard | ||
| 8 | Step-Audio-EditX | 16.5% | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | gpt-realtime | 96.7% | Aug 2025 | Official leaderboard | |
| 2 | Kimi-Audio-7B-Instruct | 34.7% | Apr 2025 | Official leaderboard |
Gpt-realtime on the official leaderboard passed 85% in August 2025
Models released before it came out already scored that high. All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official SpeechEditBench leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Audio and video
- Video-MMMU In use 300 expert-level lecture videos with questions testing perception, comprehension and adaptation of knowledge. Accuracy.
Qwen 3.8 Max 88.7%
- LVBench In use Hours-long video understanding with multiple-choice questions on reasoning, grounding and summarization.
Gemini 4 Argon 91.7%
- MMVU In use 3,000 expert-annotated questions on videos across 27 disciplines. Multiple-choice and open-ended accuracy.
Qwen 3.8 Max 82.4%
- MLVU In use Multi-task long video understanding (minutes to hours) with multiple-choice and generation tasks. Accuracy.
Qwen 3.8 Max 90.8%
- FLEURS In use Models transcribe read speech in 102 languages built from FLoRes-101 sentences. Model cards usually report word error rate, so lower scores are better.
Phi-4-multimodal 4.00%
- VoiceBench In use Spoken instructions with speaker, environment and content variation. Voice assistant responses scored by GPT judge or rules.
Qwen3.5-Omni-Plus 93.1%
Data audit
All 8 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
- Pass Independent runs Scores from the benchmark’s paper (8 entries).
- Pass Models scored 2 models we track, plus 6 other entries on its leaderboard.
- Note Newest score The newest came 15 months ago, in August 2025, for gpt-realtime. That’s expected for a saturated test.
- Note Frontier models None of its models are on the capabilities index.
- Note Headroom Saturated. The best is 96.7%, past 85% since August 2025.
Leaderboard import 3 pass
- Pass Importer We read the board on aclanthology.org: 2 rows for models we track and 6 other entries at the last read.
- Pass Last read November 21st, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.