# Automatic Speech Recognition on Chinese Dialect Scenario Aggregate

Source: [Automatic Speech Recognition on Chinese Dialect Scenario Aggregate](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-chinese-dialect-scenario-aggregate>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-10T08:51:46.632Z

Evaluation of Automatic Speech Recognition performance on the Chinese Dialect Scenario Aggregate dataset, typically measured using Hallucination rate.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Automatic Speech Recognition |
| Dataset (including any stated split) | Chinese Dialect Scenario Aggregate |
| Metric | Hallucination Rate |
| Unit | PERCENT |
| Direction | Lower is better |
| Results loaded | 8 of 8 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Hallucination Rate](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-chinese-dialect-scenario-aggregate.md?metric=Hallucination%20Rate>): Lower is better; unit: PERCENT.

## Recorded results

| Method | Hallucination Rate | Result date | Report paper |
| --- | --- | --- | --- |
| Qwen3-ASR-1.7B | 12 | Not recorded | 69d884287d0172140ec77a98 |
| Ours | 12.2 | Not recorded | 69d884287d0172140ec77a98 |
| Step-Audio2-mini | 19.4 | Not recorded | 69d884287d0172140ec77a98 |
| Ours (w/o IA-SFT) | 19.8 | Not recorded | 69d884287d0172140ec77a98 |
| GLM-ASR-nano | 20.1 | Not recorded | 69d884287d0172140ec77a98 |
| Fun-ASR-nano | 21.7 | Not recorded | 69d884287d0172140ec77a98 |
| FireRedASR-LLM | 22.8 | Not recorded | 69d884287d0172140ec77a98 |
| Qwen3-Omni-Inst | 37 | Not recorded | 69d884287d0172140ec77a98 |

### Qwen3-ASR-1.7B

Hallucination Rate: 12. Result ID: 69d8ba1c47f34ce924c1b327.

[Reporting paper](<https://www.sota2.com/research/paper/69d884287d0172140ec77a98.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76b71e209131cea6eaaf.md>)

[Implementation](<https://github.com/FireRedTeam/FireRedASR>)

Recorded evaluation configuration: {"Decoding Mode":"offline"}. This record does not assert independent reproduction.

### Ours

Hallucination Rate: 12.2. Result ID: 69d8ba1c47f34ce924c1b32c.

[Reporting paper](<https://www.sota2.com/research/paper/69d884287d0172140ec77a98.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69d884287d0172140ec77a98.md>)

Recorded evaluation configuration: {"Decoding Mode":"offline","Beam Size":"3","IA-SFT":"true"}. This record does not assert independent reproduction.

### Step-Audio2-mini

Hallucination Rate: 19.4. Result ID: 69d8ba1c47f34ce924c1b329.

[Reporting paper](<https://www.sota2.com/research/paper/69d884287d0172140ec77a98.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b881e209131cea88f3d.md>)

Recorded evaluation configuration: {"Decoding Mode":"offline"}. This record does not assert independent reproduction.

### Ours (w/o IA-SFT)

Hallucination Rate: 19.8. Result ID: 69d8ba1c47f34ce924c1b32b.

[Reporting paper](<https://www.sota2.com/research/paper/69d884287d0172140ec77a98.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69d884287d0172140ec77a98.md>)

Recorded evaluation configuration: {"Decoding Mode":"offline","Beam Size":"3","IA-SFT":"false"}. This record does not assert independent reproduction.

### GLM-ASR-nano

Hallucination Rate: 20.1. Result ID: 69d8ba1c47f34ce924c1b326.

[Reporting paper](<https://www.sota2.com/research/paper/69d884287d0172140ec77a98.md>)

[Method origin paper](<https://www.sota2.com/research/paper/697c3afda8b01e76cc8da424.md>)

Recorded evaluation configuration: {"Decoding Mode":"offline"}. This record does not assert independent reproduction.

### Fun-ASR-nano

Hallucination Rate: 21.7. Result ID: 69d8ba1c47f34ce924c1b325.

[Reporting paper](<https://www.sota2.com/research/paper/69d884287d0172140ec77a98.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7be71e209131cea8b323.md>)

Recorded evaluation configuration: {"Decoding Mode":"offline"}. This record does not assert independent reproduction.

### FireRedASR-LLM

Hallucination Rate: 22.8. Result ID: 69d8ba1c47f34ce924c1b328.

[Reporting paper](<https://www.sota2.com/research/paper/69d884287d0172140ec77a98.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b021e209131cea8635d.md>)

Recorded evaluation configuration: {"Decoding Mode":"offline"}. This record does not assert independent reproduction.

### Qwen3-Omni-Inst

Hallucination Rate: 37. Result ID: 69d8ba1c47f34ce924c1b32a.

[Reporting paper](<https://www.sota2.com/research/paper/69d884287d0172140ec77a98.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76b71e209131cea6eaaf.md>)

[Implementation](<https://github.com/FireRedTeam/FireRedASR>)

Recorded evaluation configuration: {"Decoding Mode":"offline"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
