Benchmark Almanac

ChildMandarin

Not in model cards

Models transcribe Mandarin speech from children ages three to five or verify speakers.

The publisher reports model results.

Made by
BAAI FlagEval
Released
September 2024
Kind
Speech recognition and translation
Level
Everyday
Scored
Checked automatically
Answers
Short answer
Tasks
All public
Measures
Capability
Language
Chinese
In and out
Audio → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
In views
Audio and speech
Last model added
August 2024
Scores refreshed
November 20th, 2026
16
Entries on its own leaderboard, 1 model we track

Qwen2-Audio-7B-Instruct leads with 11.5% Qwen2-Audio-7B-Instruct leads with 11.5%

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. 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. 1 Qwen2-Audio-7B-Instruct Zero-shot 11.5%
  2. 2 SenseVoice (Small) Zero-shot 11.9%
  3. 3 CW Fine-tuning 13.7%
  4. 4 HuBERT (L) Greedy search 15%
  5. 5 Whisper-small Fine-tuning 17.4%
  6. 6 HuBERT (B) Greedy search 18.7%
  7. 7 Whisper-medium Fine-tuning 19%
  8. 8 Wav2vec 2.0 (B) Greedy search 20.3%
  9. 9 Qwen-Audio Zero-shot 20.4%
  10. 10 Wav2vec 2.0 (L) Greedy search 21.1%
  11. 11 Whisper-base Fine-tuning 23.3%
  12. 12 Conformer CTC+AED, Attention Rescoring 27.4%
  13. 13 Whisper-tiny Fine-tuning 28.8%
  14. 14 Paraformer Paraformer, Beam 28.9%
  15. 15 Whisper-large-v2 Zero-shot 29.4%
  16. 16 Transformer CTC+AED, Attention Rescoring 32.1%

Source: Official ChildMandarin leaderboard

Source: Official ChildMandarin leaderboard

#EntryScoreSettingDateSource
1 Qwen2-Audio-7B-Instruct 11.5% Zero-shot Aug 2024 Official leaderboard
2 SenseVoice (Small) 0.2B 11.9% Zero-shot Official leaderboard
3 CW 0.1B 13.7% Fine-tuning Official leaderboard
4 HuBERT (L) 15% Greedy search Official leaderboard
5 Whisper-small 0.2B 17.4% Fine-tuning Official leaderboard
6 HuBERT (B) 18.7% Greedy search Official leaderboard
7 Whisper-medium 0.8B 19% Fine-tuning Official leaderboard
8 Wav2vec 2.0 (B) 20.3% Greedy search Official leaderboard
9 Qwen-Audio 7.7B 20.4% Zero-shot Official leaderboard
10 Wav2vec 2.0 (L) 21.1% Greedy search Official leaderboard
11 Whisper-base 0.1B 23.3% Fine-tuning Official leaderboard
12 Conformer 0B 27.4% CTC+AED, Attention Rescoring Official leaderboard
13 Whisper-tiny 0B 28.8% Fine-tuning Official leaderboard
14 Paraformer 0B 28.9% Paraformer, Beam Official leaderboard
15 Whisper-large-v2 1.6B 29.4% Zero-shot Official leaderboard
16 Transformer 0B 32.1% CTC+AED, Attention Rescoring Official leaderboard
#ModelScoreSettingDateSource
1 Qwen2-Audio-7B-Instruct 11.5% Zero-shot Aug 2024 Official leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard
0% 20% 40% 60% 80% 100% Qwen2-Audio-7B-Instruct 0% 20% 40% 60% 80% 100% Qwen2-Audio-7B-Instruct

Source: Official ChildMandarin 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

All 695 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

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 2 years ago, in August 2024, for Qwen2-Audio-7B-Instruct. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (16 entries).
  • Pass Models scored 1 model we track, plus 15 other entries on its leaderboard.
  • Warning Newest score The newest came 2 years ago, in August 2024, for Qwen2-Audio-7B-Instruct.
  • Note Frontier models None of its models are on the capabilities index.
  • Note Headroom Lower is better, so there’s no ceiling to reach.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 1 row for models we track and 15 other entries at the last read.
  • Pass Last read November 20th, 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.

How we check