Afrispeech Semantics - Entailment (AfriSpeech-200): leaderboard

Metric: Macro-F1 (%; three-way entailment, neutral or contradiction judgments of text hypotheses against a spoken premise, computed over all premise-hypothesis pairs of the AfriSpeech-200 set; zero-shot generative audio language model, free-form answers mapped to labels by a lightweight language model). Source: arxiv.org. Saturation forecast: Around December 2026. 7 models tracked.

Top models

#ModelScore
1Qwen2-Audio-7B-Instruct71.23
2Qwen2.5-Omni-7B67.31
3SALMONN-13B28.14

Interactive version: theaggregate.ai/benchmark?slug=afrispeech-semantics-entailment-afrispeech-200 · How It Works · Data refreshed daily, snapshot 2026-09-26.