SEA-SpeechBench - Age Recognition (SEA Prompt): leaderboard

Metric: Macro-F1 (%; speaker age group (teens, adults, seniors) from speech, labels canonicalised by bilingual string matching, per-dataset scores averaged; short clips of at most 30 s from curated public Southeast Asian speech corpora, up to 1,000 sampled items per dataset; instruction prompt in the native Southeast Asian language). Source: arxiv.org. Saturation forecast: Around 2028. 15 models tracked.

Top models

#ModelScore
1Qwen3 Omni 30B A3B Instruct36.7
2Gemini 2.5 Flash36.33
3GPT-4o Audio34.27
4Gemma 3n E4B (IT)29.97
5Voxtral-Mini-3B-250728.49
6Phi-4 Multimodal Instruct28.42
7gemma-3n-E2B-it22.7
8Qwen2.5-Omni-7B18.28
9SeaLLMs-Audio-7B18.08
10Qwen2-Audio-7B-Instruct15.39

Interactive version: theaggregate.ai/benchmark?slug=sea-speechbench-age-recognition-sea-prompt · How It Works · Data refreshed daily, snapshot 2026-09-26.