MHGraphBench - Entity Typing: leaderboard
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.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 Mini | 98.48 |
| 2 | GPT-5.1 Instant | 98.27 |
| 3 | GPT-5.2 Instant | 98.05 |
| 4 | GPT-4.1 | 97.62 |
| 5 | GPT-4o | 97.4 |
| 6 | Qwen 2.5 32B Instruct | 76.45 |
| 7 | Mistral 7B Instruct (v0.3) | 70.76 |
| 8 | Qwen 2.5 7B Instruct | 49.32 |
| 9 | DeepSeek-R1-Distill-Qwen-7B | 9.42 |
| 10 | Llama 3.1 8B Instruct | 8.39 |
| 11 | DeepSeek R1 Distill Qwen 32B | 6.55 |
Interactive version: theaggregate.ai/benchmark?slug=mhgraphbench-entity-typing · How It Works · Data refreshed daily, snapshot 2026-10-07.