AEQ-Bench - Naturalness: leaderboard

Metric: Naturalness (1-4; how human-like and conversational the text response is, from analysis only (1) to human-like chatting (4), rated by a GPT-5-mini judge with manual checking; mean over the Meld, GigaSpeech and EmoV-DB subsets, 1,885 instances). Source: arxiv.org. Saturation forecast: Around November 2026. 9 models tracked.

Top models

#ModelScore
1Qwen2.5-Omni-7B3.89
2Qwen3 Omni 30B A3B Instruct3.77
3SALMONN-7B3.24
4Qwen2-Audio-7B-Instruct2.82
5Qwen-Audio-Chat2.37

Interactive version: theaggregate.ai/benchmark?slug=aeq-bench-naturalness · How It Works · Data refreshed daily, snapshot 2026-09-26.