MHGraphBench - Two-hop Selection: leaderboard

Two-hop selection questions over drug-disease-disease paths in MHGraphBench (2026), a knowledge-graph-grounded mental-health benchmark built from a curated PrimeKG subgraph, with letter-only multiple-choice answers. Accuracy in percent; higher is better.

Metric: Accuracy (%) on the 1,200 two-hop selection items, letter-only answers; the OpenAI API models answer at temperature 0 (at most 120 completion tokens) with strict answer-letter parsing, the open models are scored by forced-choice option-letter log-probabilities; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 15 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-5.2 Instant65.58#304
2GPT-4.164#505
3GPT-5.1 Instant60.83#472
4GPT-5 Mini59.83#364
5Qwen 2.5 32B Instruct59#1247
6GPT-4o54.25#751
7Qwen 2.5 7B Instruct38.92#2723
8DeepSeek R1 Distill Qwen 32B33.92#1834
9Llama 3.1 8B Instruct33.83#4600
10DeepSeek-R1-Distill-Qwen-7B27.17#8055
11Mistral 7B Instruct (v0.3)26.17#6750
12Llama3-Med42-8B24.92#5084
13Llama3-OpenBioLLM-8B24.58#6364
14BioMistral-7B22.92#8332
15Meditron-7B21.42#8733

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

Interactive version: theaggregate.ai/benchmark?slug=mhgraphbench-two-hop-selection · How It Works · Data refreshed daily, snapshot 2026-11-21.