Benchmark
AHELM: Robust Speech Bench - WER
Scores are percentages. Lower is better.
- Models measured
- 12
- Scores recorded
- 132
- Sources
- 1
- IQ weight
- 1.49
Score distribution
- 0.287%
- Mean
- 0.103%
- Median
- 0.453%
- Std dev
- 0.044%
- 10th percentile
- 0.724%
- 90th percentile
- 0.0387%
- Min
This benchmark has a weight of 1.49 for chat models, measured over 9 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro Preview 05-06subject: librispeech_env_noise, level: 1· AHELM (HELM Audio) data | 0.0387% |
| 2 | Gemini 2.5 Pro Preview 05-06subject: librispeech_env_noise, level: 2· AHELM (HELM Audio) data | 0.0387% |
| 3 | Gemini 2.5 Pro Preview 05-06subject: librispeech_env_noise, level: 3· AHELM (HELM Audio) data | 0.0387% |
| 4 | Gemini 2.5 Pro Preview 05-06· AHELM (HELM Audio) data | 0.0388% |
| 5 | Gemini 2.5 Pro Preview 05-06subject: librispeech_gnoise, level: 1· AHELM (HELM Audio) data | 0.0388% |
| 6 | Gemini 2.5 Pro Preview 05-06subject: librispeech_gnoise, level: 2· AHELM (HELM Audio) data | 0.0388% |
| 7 | Gemini 2.5 Pro Preview 05-06subject: librispeech_gnoise, level: 3· AHELM (HELM Audio) data | 0.0388% |
| 8 | Gemini 2.0 Flashsubject: librispeech_env_noise, level: 2· AHELM (HELM Audio) data | 0.0435% |
| 9 | Gemini 2.0 Flashsubject: librispeech_env_noise, level: 3· AHELM (HELM Audio) data | 0.0435% |
| 10 | Gemini 2.0 Flashsubject: librispeech_env_noise, level: 1· AHELM (HELM Audio) data | 0.0436% |
| 11 | Gemini 2.0 Flashsubject: librispeech_gnoise, level: 1· AHELM (HELM Audio) data | 0.0437% |
| 12 | Gemini 2.0 Flashsubject: librispeech_gnoise, level: 2· AHELM (HELM Audio) data | 0.0438% |
| 13 | Gemini 2.0 Flashsubject: librispeech_gnoise, level: 3· AHELM (HELM Audio) data | 0.0439% |
| 14 | Gemini 2.0 Flash Experimentalsubject: librispeech_env_noise, level: 1· AHELM (HELM Audio) data | 0.0443% |
| 15 | Gemini 2.0 Flash Experimentalsubject: librispeech_env_noise, level: 3· AHELM (HELM Audio) data | 0.0443% |
| 16 | Gemini 2.0 Flash Experimentalsubject: librispeech_gnoise, level: 1· AHELM (HELM Audio) data | 0.0444% |
| 17 | Gemini 2.0 Flash Experimentalsubject: librispeech_gnoise, level: 2· AHELM (HELM Audio) data | 0.0444% |
| 18 | Gemini 2.0 Flash Experimentalsubject: librispeech_gnoise, level: 3· AHELM (HELM Audio) data | 0.0444% |
| 19 | Gemini 2.0 Flash Experimentalsubject: librispeech_env_noise, level: 2· AHELM (HELM Audio) data | 0.0446% |
| 20 | Gemini 2.0 Flash Litesubject: librispeech_env_noise, level: 2· AHELM (HELM Audio) data | 0.0491% |
| 21 | Gemini 2.0 Flash Litesubject: librispeech_gnoise, level: 1· AHELM (HELM Audio) data | 0.0492% |
| 22 | Gemini 2.0 Flash Litesubject: librispeech_gnoise, level: 2· AHELM (HELM Audio) data | 0.0492% |
| 23 | Gemini 2.0 Flash Litesubject: librispeech_gnoise, level: 3· AHELM (HELM Audio) data | 0.0492% |
| 24 | Gemini 2.0 Flash Lite· AHELM (HELM Audio) data | 0.0492% |
| 25 | Gemini 2.0 Flash Litesubject: librispeech_env_noise, level: 1· AHELM (HELM Audio) data | 0.0492% |
| 26 | Gemini 2.0 Flash Litesubject: librispeech_env_noise, level: 3· AHELM (HELM Audio) data | 0.0492% |
| 27 | gemini-1.5-pro-002subject: librispeech_gnoise, level: 3· AHELM (HELM Audio) data | 0.0528% |
| 28 | gemini-1.5-pro-002subject: librispeech_gnoise, level: 2· AHELM (HELM Audio) data | 0.0529% |
| 29 | gemini-1.5-pro-002subject: librispeech_gnoise, level: 1· AHELM (HELM Audio) data | 0.0536% |
| 30 | gemini-1.5-pro-002subject: librispeech_env_noise, level: 2· AHELM (HELM Audio) data | 0.0549% |
| 31 | gemini-1.5-pro-002subject: librispeech_env_noise, level: 3· AHELM (HELM Audio) data | 0.0549% |
| 32 | gemini-1.5-pro-002subject: librispeech_env_noise, level: 1· AHELM (HELM Audio) data | 0.0551% |
| 33 | gemini-1.5-flash-002subject: librispeech_env_noise, level: 1· AHELM (HELM Audio) data | 0.0617% |
| 34 | gemini-1.5-flash-002subject: librispeech_env_noise, level: 2· AHELM (HELM Audio) data | 0.0619% |
| 35 | Gemini 2.0 Flashsubject: accented_cv_es· AHELM (HELM Audio) data | 0.0627% |
| 36 | gemini-1.5-flash-002subject: librispeech_env_noise, level: 3· AHELM (HELM Audio) data | 0.0628% |
| 37 | gemini-1.5-flash-002subject: librispeech_gnoise, level: 1· AHELM (HELM Audio) data | 0.0631% |
| 38 | gemini-1.5-flash-002subject: librispeech_gnoise, level: 2· AHELM (HELM Audio) data | 0.0631% |
| 39 | gemini-1.5-flash-002subject: librispeech_gnoise, level: 3· AHELM (HELM Audio) data | 0.0631% |
| 40 | gemini-1.5-pro-002subject: accented_cv_es· AHELM (HELM Audio) data | 0.0634% |
| 41 | Gemini 2.0 Flash Experimentalsubject: accented_cv_es· AHELM (HELM Audio) data | 0.0636% |
| 42 | gemini-1.5-pro-001subject: librispeech_gnoise, level: 1· AHELM (HELM Audio) data | 0.0693% |
| 43 | gemini-1.5-pro-001subject: librispeech_env_noise, level: 3· AHELM (HELM Audio) data | 0.0697% |
| 44 | gemini-1.5-pro-001subject: librispeech_gnoise, level: 2· AHELM (HELM Audio) data | 0.07% |
| 45 | gemini-1.5-pro-001subject: librispeech_env_noise, level: 2· AHELM (HELM Audio) data | 0.07% |
| 46 | gemini-1.5-flash-002subject: accented_cv_es· AHELM (HELM Audio) data | 0.0702% |
| 47 | gemini-1.5-pro-001subject: librispeech_env_noise, level: 1· AHELM (HELM Audio) data | 0.0705% |
| 48 | gemini-1.5-pro-001subject: librispeech_gnoise, level: 3· AHELM (HELM Audio) data | 0.0707% |
| 49 | gemini-1.5-pro-001subject: accented_cv_es· AHELM (HELM Audio) data | 0.0758% |
| 50 | Gemini 2.5 Flash Preview 05-20 (no think)· AHELM (HELM Audio) data | 0.0766% |
Showing 1–50 of 132 results.