AISHELL-2
Models transcribe 1,000 hours of Mandarin speech recorded on several devices. Scored by character error rate.
A larger corpus than AISHELL-1 with its own test sets, so scores are not comparable.
In the authors’ words
AISHELL-1 is by far the largest open-source speech corpus available for Mandarin speech recognition research. It was released with a baseline system containing solid training and testing pipelines for Mandarin ASR. In AISHELL-2, 1000 hours of clean read-speech data from iOS is published, which is free for academic usage. On top of AISHELL-2 corpus, an improved recipe is developed and released, containing key components for industrial applications, such as Chinese word segmentation, flexible vocabulary expension and phone set transformation etc. Pipelines support various state-of-the-art techniques, such as time-delayed neural networks and Lattic-Free MMI objective funciton. In addition, we also release dev and test data from other channels(Android and Mic). For research community, we hope that AISHELL-2 corpus can be a solid resource for topics like transfer learning and robust ASR. For industry, we hope AISHELL-2 recipe can be a helpful reference for building meaningful industrial systems and products.
- Level
- Everyday Most adults, without training
- Scored
- Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
- Answers
- Short answer A number, word or short string matched to a key.
Kimi-Audio-7B-Instruct leads with 2.56% among models released by Jan 2026
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
The newest model with a score here came out in Jan 2026, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6 Astra
- Claude Opus 5.5
- GPT-6.1 Terra
- GPT-6.1 Sol
- and 16 more
- 1 Kimi-Audio-7B-Instruct 2.56%
- 2
ERNIE 5.0 2.64% - 3
LongCat-Flash-Omni 2.78%
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Kimi-Audio-7B-Instruct | 2.56% | Apr 2025 | Model card | |
| 2 | ERNIE 5.0 | 2.64% | Feb 2026 | Model card | |
| 3 | LongCat-Flash-Omni | 2.78% | Oct 2025 | Model card |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
No launch post has listed it since Oct 2025
Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.
| Date | Launch post | Score | As the post writes it |
|---|---|---|---|
| 31 Oct 2025 | LongCat-Flash-Omni | 2.78 | AISHELL-2 |
| 25 Apr 2025 | Kimi-Audio-7B-Instruct | 2.56 | AISHELL-2 ios |
3 system cards report it, from 3 labs
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| 4 Feb 2026 | ERNIE 5.0 technical report | 2.64 | AISHELL-2, WER, lower is better (Table 7) | Capabilities |
| 31 Oct 2025 | LongCat-Flash-Omni technical report | 2.78 | AISHELL-2, CER, lower is better (Table 9) | Capabilities |
| 25 Apr 2025 | Kimi-Audio-7B-Instruct technical report | 2.56 | AISHELL-2 ios, WER, lower is better (Table 4) | Capabilities |
Versions
We switch the top row to each new version when it comes out. Open it for one row per version. An old version stays in play while people still test it, so whoever first takes it past 85% gets credit for closing it out.
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%
- 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%
- TVBench In use Models answer multiple-choice questions on 10 temporal video tasks, such as action counting and ordering, that static frames cannot solve. Scored by accuracy. Qwen 3.8 Max 81.9%