Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Spoken Question Answering research benchmarks – Page 2 · SOTA2 Research

Spoken Question Answering

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

66.62Accuracy

8

May 26, 2026

76.34Accuracy

8

Feb 26, 2026

83.95Accuracy

8

Feb 26, 2026

18.54Accuracy

8

Feb 26, 2026

3.88WER

7

May 12, 2026

Llama Questions

Qwen2.5-Omni-7B

0.99WER

7

Mar 11, 2026

49.9PEDANTS Score

6

Jun 11, 2026

40.8Text Score

6

May 12, 2026

45.1Text Score

6

May 12, 2026

74.2Text Score

6

May 12, 2026

Spoken SQuAD NoiseV2 (test)

BiDAF (WORD+CHAR+PHONEME+SYLLABLE)

20.66EM

5

Feb 26, 2026

Spoken SQuAD NoiseV1 (test)

BiDAF (WORD+CHAR+PHONEME+SYLLABLE)

29.73EM

5

Feb 26, 2026

Spoken SQuAD No noise (test)

BiDAF (WORD+CHAR+PHONEME+SYLLABLE)

45.78EM

5

Feb 26, 2026

LibriSQA Part I and Part II

SpeechMapper + LoRA (C)

42.8MCIF

4

Jul 3, 2026

1,750N Verified Verifications Count

4

Jun 17, 2026

Longcat-Flash-Omni-Instruct

4.94AlpacaEval

4

Feb 26, 2026

OpenAudioBench S2T

Fun-Audio-Chat-30B-A3B

88.89AlpacaEval Score

4

Feb 26, 2026

Web Questions N=2032 (test)

18.1Accuracy

3

Jun 15, 2026

Llama Questions N=300 (test)

46Accuracy

3

Jun 15, 2026

31.12BLEU

3

Jun 9, 2026

Spoken SQuAD (A) Text (test)

76.9EM

3

Feb 26, 2026

SQuAD lost (test)

37.31F1 Score

2

Feb 26, 2026

Page 2 of 2

Popular datasets

Spoken SQuAD LibriSQA SQuAD TriviaQA UltraEval-Audio

Follow for updates

Join SOTA2 Contact