AMI-ME - Meeting Effectiveness Scoring: leaderboard

AMI-ME (2026) evaluates automatic meeting effectiveness scoring: models rate the effectiveness of fine-grained topic segments of AMI Corpus meetings against their objectives. Segment-level Spearman correlation with professional annotators, from -1 to 1; higher is better.

Metric: Segment-level Spearman correlation (-1 to 1) between the model's predicted effectiveness scores and the mean of three human annotators' scores on AMI-ME (all segments of 130 AMI Corpus meetings), given gold topic segments and meeting objectives; G-Eval style scoring (probability-weighted scores, or the mean of five samples for reasoning models), context window 1; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 5 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Qwen 3 32BNon-reasoning0.6445#972
2GPT-4o0.6341#751
3DeepSeek R1 Distill Llama 70B0.6132#1669
4Llama 3.3 70B Instruct0.6072#1385
5Gemini 2.5 Flash0.5624#463

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=ami-me-meeting-effectiveness-scoring · How It Works · Data refreshed daily, snapshot 2026-11-21.