MHGraphBench - Entity Typing: leaderboard

Five-way questions on the type of an entity 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 1,847 five-way entity typing 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: Estimated already saturated. 15 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-5 Mini98.48#364
2GPT-5.1 Instant98.27#472
3GPT-5.2 Instant98.05#304
4GPT-4.197.62#505
5GPT-4o97.4#751
6Qwen 2.5 32B Instruct76.45#1247
7Mistral 7B Instruct (v0.3)70.76#6750
8Qwen 2.5 7B Instruct49.32#2723
9BioMistral-7B29.24#8332
10Llama3-Med42-8B24.96#5084
11Meditron-7B20.25#8733
12Llama3-OpenBioLLM-8B16.57#6364
13DeepSeek-R1-Distill-Qwen-7B9.42#8055
14Llama 3.1 8B Instruct8.39#4600
15DeepSeek R1 Distill Qwen 32B6.55#1834

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

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