Skip to content

Benchmark

AHELM: MuTox - EM

Scores are percentages.

Models measured
12
Scores recorded
359
Sources
1
IQ weight
No weight published

Score distribution

75.5%
Mean
77.9%
Median
15.6%
Std dev
53.4%
10th percentile
100%
90th percentile
33.3%
Min
33.3%100%
The band covers the middle 80% of scores. The line marks the median.

Leaderboard

#ModelScore
51Gemini 2.5 Flash Preview 05-20 (no think)language: Catalan· AHELM (HELM Audio) data91.4%
52Gemini 2.0 Flashlanguage: Tagalog· AHELM (HELM Audio) data90.9%
53Gemini 2.0 Flash Experimentallanguage: Tagalog· AHELM (HELM Audio) data90.9%
54Gemini 2.5 Flash Preview 05-20 (no think)language: Tagalog· AHELM (HELM Audio) data90.9%
55gemini-1.5-flash-002language: Tagalog· AHELM (HELM Audio) data90.9%
56GPT-4o Audio Previewlanguage: Tagalog· AHELM (HELM Audio) data90.9%
57Qwen2.5-Omni 7Blanguage: Tagalog· AHELM (HELM Audio) data90.9%
58GPT-4o Audio Previewlanguage: Finnish· AHELM (HELM Audio) data90.8%
59GPT-4o mini Audio Previewlanguage: Finnish· AHELM (HELM Audio) data90.8%
60GPT-4o mini Audio Previewlanguage: Slovak· AHELM (HELM Audio) data90.8%
61GPT-4o Audio Previewlanguage: Greek· AHELM (HELM Audio) data90.5%
62GPT-4o Audio Previewlanguage: Swahili· AHELM (HELM Audio) data90%
63GPT-4o mini Audio Previewlanguage: Swahili· AHELM (HELM Audio) data90%
64Qwen2.5-Omni 7Blanguage: Swahili· AHELM (HELM Audio) data90%
65Qwen2.5-Omni 7Blanguage: Slovak· AHELM (HELM Audio) data89.6%
66Qwen2.5-Omni 7Blanguage: Finnish· AHELM (HELM Audio) data89.6%
67GPT-4o mini Audio Previewlanguage: Polish· AHELM (HELM Audio) data89.3%
68GPT-4o Audio Previewlanguage: Mandarin_Chinese· AHELM (HELM Audio) data88.9%
69GPT-4o mini Audio Previewlanguage: Mandarin_Chinese· AHELM (HELM Audio) data88.9%
70gemini-1.5-flash-002language: Bengali· AHELM (HELM Audio) data88.2%
71gemini-1.5-pro-002language: Bengali· AHELM (HELM Audio) data88.2%
72GPT-4o Audio Previewlanguage: Bengali· AHELM (HELM Audio) data88.2%
73GPT-4o mini Audio Previewlanguage: Bengali· AHELM (HELM Audio) data88.2%
74Gemini 2.5 Pro Preview 05-06language: Catalan· AHELM (HELM Audio) data87.6%
75Gemini 2.0 Flashlanguage: French· AHELM (HELM Audio) data87.5%
76Gemini 2.0 Flash Experimentallanguage: French· AHELM (HELM Audio) data87.5%
77GPT-4o mini Audio Preview· AHELM (HELM Audio) data87.4%
78Gemini 2.0 Flash Litelanguage: Estonian· AHELM (HELM Audio) data87.3%
79Gemini 2.5 Flash Preview 05-20 (no think)language: Estonian· AHELM (HELM Audio) data87.3%
80GPT-4o mini Audio Previewlanguage: Greek· AHELM (HELM Audio) data87.2%
81Gemini 2.5 Pro Preview 05-06language: Estonian· AHELM (HELM Audio) data86.7%
82Qwen2.5-Omni 7Blanguage: Catalan· AHELM (HELM Audio) data86.5%
83GPT-4o Audio Previewlanguage: Polish· AHELM (HELM Audio) data86.4%
84GPT-4o mini Audio Previewlanguage: Hebrew· AHELM (HELM Audio) data86.2%
85gemini-1.5-pro-002language: Estonian· AHELM (HELM Audio) data86.1%
86Gemini 2.0 Flash Litelanguage: Bulgarian· AHELM (HELM Audio) data85.9%
87GPT-4o Audio Preview· AHELM (HELM Audio) data85.9%
88Gemini 2.0 Flash Litelanguage: Finnish· AHELM (HELM Audio) data85.9%
89GPT-4o Audio Previewlanguage: Czech· AHELM (HELM Audio) data85.8%
90Qwen2.5-Omni 7Blanguage: Greek· AHELM (HELM Audio) data85.8%
91Gemini 2.0 Flash Litelanguage: Hindi· AHELM (HELM Audio) data85.7%
92Gemini 2.0 Flash Litelanguage: Turkish· AHELM (HELM Audio) data85.7%
93Gemini 2.0 Flash Litelanguage: Urdu· AHELM (HELM Audio) data85.7%
94Gemini 2.5 Flash Preview 05-20 (no think)language: Hindi· AHELM (HELM Audio) data85.7%
95Gemini 2.5 Flash Preview 05-20 (no think)language: Turkish· AHELM (HELM Audio) data85.7%
96Gemini 2.5 Pro Preview 05-06language: Turkish· AHELM (HELM Audio) data85.7%
97gemini-1.5-flash-001language: Turkish· AHELM (HELM Audio) data85.7%
98gemini-1.5-flash-001language: Urdu· AHELM (HELM Audio) data85.7%
99gemini-1.5-flash-002language: Turkish· AHELM (HELM Audio) data85.7%
100gemini-1.5-flash-002language: Urdu· AHELM (HELM Audio) data85.7%

Showing 51–100 of 359 results.