Afrispeech Semantics - Consistency (AfriSpeech-General): leaderboard

Metric: Macro-F1 (%; binary consistent versus inconsistent judgments of text hypotheses against a spoken premise, computed over all premise-hypothesis pairs of the AfriSpeech-General set; zero-shot generative audio language model, free-form answers mapped to labels by a lightweight language model). Source: arxiv.org. Saturation forecast: Estimated already saturated. 7 models tracked.

Top models

#ModelScore
1Qwen2.5-Omni-7B91.31
2SALMONN-13B90.52
3Qwen2-Audio-7B-Instruct82.33

Interactive version: theaggregate.ai/benchmark?slug=afrispeech-semantics-consistency-afrispeech-general · How It Works · Data refreshed daily, snapshot 2026-09-26.