Deep-Research Citation Benchmark - Source Relevance: leaderboard

Metric: Cohen kappa (-1 to 1) between the LLM judge and human-reviewed gold labels on the source relevance rubric criterion of 624 attribution-citation pairs parsed from an adversarial 25-domain deep-research report (scores binarized at 0.5); higher is better. Source: arxiv.org. Saturation forecast: Around February 2028. 8 models tracked.

Top models

#ModelScore
1GPT-5 Mini0.64
2Claude Opus 4.60.55
3GPT-5.4 Mini0.54
4GPT-OSS-120B0.52
5Gemini 3.1 Pro (Preview)0.46
6Gemini 3.1 Flash Lite (Preview)0.46
7Claude Haiku 4.50.34
8Claude Sonnet 4.60.32

Interactive version: theaggregate.ai/benchmark?slug=deep-research-citation-benchmark-source-relevance · How It Works · Data refreshed daily, snapshot 2026-09-29.