ReSUM - SQuALITY - chrF: leaderboard
ReSUM v1 Table5: query-focused long-meeting summarization on QMSum and SQuALITY; source-native zero-shot, CoT, RAG, fine-tuned and multi-agent methods remain separate run identities. Short source model labels (including LLaMA3-3B and GPT-4-mini) are not bound to a guessed checkpoint. R1/R2/RL, BERTScore, Mover and COMET are native 0–1 values; chrF is native percent. No transformation, model alias, cross-method pooling or resource metric.
Metric: Score (%). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 16 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Our Method (ReSUM Run: method=ReSUM Multi-Agent;checkpoint=source-native;effort=unreported) | 43.78 | |
| 2 | GPT-4-mini (ReSUM Run: method=Zero-shot;checkpoint=source-native;effort=unreported) | 39.34 | |
| 3 | FT+Res Qwen (ReSUM Run: method=End-to-End Fine-tuned with Reasoning;checkpoint=source-native;effort=unreported) | 38 | |
| 4 | Gemma3-4B (ReSUM Run: method=Zero-shot;checkpoint=source-native;effort=unreported) | 37.38 | |
| 5 | FT+Res LLaMA (ReSUM Run: method=End-to-End Fine-tuned with Reasoning;checkpoint=source-native;effort=unreported) | 35.74 | |
| 6 | LLaMA3-3B (ReSUM Run: method=Zero-shot;checkpoint=source-native;effort=unreported) | 35.52 | |
| 7 | FT Qwen (ReSUM Run: method=Fine-tuned (FT);checkpoint=source-native;effort=unreported) | 34.63 | |
| 8 | COT Qwen (ReSUM Run: method=Chain-of-Thought (CoT);checkpoint=source-native;effort=unreported) | 34.49 | |
| 9 | COT LLaMA (ReSUM Run: method=Chain-of-Thought (CoT);checkpoint=source-native;effort=unreported) | 32.56 | |
| 10 | Qwen3-4B (ReSUM Run: method=Zero-shot;checkpoint=source-native;effort=unreported) | 32.35 |
Interactive version: theaggregate.ai/benchmark?slug=resum-squality-chrf · How It Works · Data refreshed daily, snapshot 2026-11-12.