Benchmark Almanac

EmotionTalk

Hubert-Large 71.3 tops its leaderboard chatglm2-6b 50.9 leads the models we track
Not in model cards

Models infer emotions and sentiment from Chinese conversations.

The publisher reports model results.

Made by
BAAI FlagEval
Released
May 2025
Kind
Audio understanding and voice
Level
Everyday
Scored
Checked automatically
Answers
Prediction
Tasks
All public
Measures
Capability
Language
Chinese
In and out
Text, Audio, Video → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
chatglm2-6b #244 on the capability index
Last model added
June 2023
Scores refreshed
November 20th, 2026
26
Entries on its own leaderboard, 1 model we track

Hubert-Large tops the leaderboard with 71.3 chatglm2-6b leads with 50.9

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Published mean on Unimodal Emotion Recognition for the specified input modality. 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. The best model we track is chatglm2-6b, at 50.9. This view shows only models we track, each at its best.

  1. 1 Hubert-Large Speech 71.3
  2. 2 Hubert-Base Speech 71.0
  3. 3 CLIP-Large Visual 67.7
  4. 4 DINOv2-Giant Visual 64.9
  5. 5 WavLM-Large Speech 64.2
  6. 6 DINOv2-Large Visual 63.8
  7. 7 Wav2vec 2.0-Large Speech 63.8
  8. 8 Wav2vec 2.0-Base Speech 63.6
  9. 9 WavLM-Base Speech 62.1
  10. 10 Whisper-Large Speech 61.8
  11. 11 CLIP-Base Visual 61.0
  12. 12 Whisper-Base Speech 59.1
  13. 13 VideoMAE-Large Visual 58.3
  14. 14 EVA-02-Base Visual 55.5
  15. 15 Baichuan-7B Text 51.6
  16. 16 RoBERTa-Base Text 51.2
  17. 17 RoBERTa-Large Text 51.1
  18. 18 chatglm2-6b Text 50.9
  19. 19 VideoMAE-Base Visual 50.6
  20. 20 BLOOM-7B Text 50.5
  21. 21 Sentence-BERT Text 50.0
  22. 22 BERT-Base Text 49.9
  23. 23 DeBERTa-Large Text 49.1
  24. 24 LERT-Base Text 48.9
  25. 25 Vicuna-7B Text 47.8
  26. 26 Datal2vec-Base Visual 34.7

Source: Official EmotionTalk leaderboard

Source: Official EmotionTalk leaderboard

#EntryScoreSettingDateSource
1 Hubert-Large 71.3 Speech Official leaderboard
2 Hubert-Base 71.0 Speech Official leaderboard
3 CLIP-Large 67.7 Visual Official leaderboard
4 DINOv2-Giant 64.9 Visual Official leaderboard
5 WavLM-Large 64.2 Speech Official leaderboard
6 DINOv2-Large 63.8 Visual Official leaderboard
7 Wav2vec 2.0-Large 63.8 Speech Official leaderboard
8 Wav2vec 2.0-Base 63.6 Speech Official leaderboard
9 WavLM-Base 62.1 Speech Official leaderboard
10 Whisper-Large 61.8 Speech Official leaderboard
11 CLIP-Base 61.0 Visual Official leaderboard
12 Whisper-Base 59.1 Speech Official leaderboard
13 VideoMAE-Large 58.3 Visual Official leaderboard
14 EVA-02-Base 55.5 Visual Official leaderboard
15 Baichuan-7B 51.6 Text Official leaderboard
16 RoBERTa-Base 51.2 Text Official leaderboard
17 RoBERTa-Large 51.1 Text Official leaderboard
18 chatglm2-6b 50.9 Text Jun 2023 Official leaderboard
19 VideoMAE-Base 50.6 Visual Official leaderboard
20 BLOOM-7B 50.5 Text Official leaderboard
21 Sentence-BERT 50.0 Text Official leaderboard
22 BERT-Base 49.9 Text Official leaderboard
23 DeBERTa-Large 49.1 Text Official leaderboard
24 LERT-Base 48.9 Text Official leaderboard
25 Vicuna-7B 47.8 Text Official leaderboard
26 Datal2vec-Base 34.7 Visual Official leaderboard
#ModelScoreSettingDateSource
1 chatglm2-6b 50.9 Text Jun 2023 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
50.9 2023 50.9 2023

Source: Official EmotionTalk 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 3 years ago, in June 2023, for chatglm2-6b. 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 (26 entries).
  • Pass Models scored 1 model we track, plus 25 other entries on its leaderboard.
  • Warning Newest score The newest came 3 years ago, in June 2023, for chatglm2-6b.
  • Note Frontier models None of the top 10 has a score. The strongest tested is chatglm2-6b (#244).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 1 row for models we track and 25 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