Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Knowledge benchmark
SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines. Every model the catalog carries a published SuperGPQA value for, ranked by that value.
An expanded version of GPQA that evaluates graduate-level knowledge and reasoning capabilities across 285 disciplines, providing comprehensive coverage of academic domains. 22 models in the ModelScale catalog have a published value.
Each point is a model with a published SuperGPQA value and complete token prices, on the 75% input / 25% output blend. Frontier models are spaced evenly along the cost axis.
22 models with a published SuperGPQA value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Multiple choice questions |
|---|---|---|---|
| 1 | Anthropic | 95 | |
| 1 | Anthropic | 95 | |
| 3 | Alibaba | 73.9 |
What the benchmark measures and how its value is published, from the source’s own record.
SuperGPQA (SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines) is a knowledge benchmark published by SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines. An expanded version of GPQA that evaluates graduate-level knowledge and reasoning capabilities across 285 disciplines, providing comprehensive coverage of academic domains. Its evaluation set is described by the source as 285 disciplines. The source rates its difficulty as graduate level.
Each model's result is published by SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines as multiple choice questions. ModelScale prints that value exactly as published, and never converts or rescales it. This page ranks the 22 models that carry a published value, highest value first; a model the source never scored is left out rather than ranked at zero. The source does not state whether a higher value is the better result, so neither does this page.
| 4 | Alibaba | 73.6 |
| 5 | Alibaba | 71.6 |
| 6 | Alibaba | 71.4 |
| 7 | ByteDance | 70.8 |
| 8 | Anthropic | 70.6 |
| 9 | Alibaba | 70.4 |
| 10 | Moonshot AI | 69.2 |
| 11 | ByteDance | 67.4 |
| 12 | Alibaba | 67.1 |
| 13 | Z.AI | 66.8 |
| 14 | Alibaba | 66 |
| 15 | Alibaba | 65.6 |
| 16 | Alibaba | 64.9 |
| 17 | Alibaba | 64.7 |
| 18 | Alibaba | 63.4 |
| 19 | Alibaba | 62.6 |
| 20 | Moonshot AI | 57.2 |
| 21 | OPMiniCPM5-2B | OpenBMB | 40.8 |
| 22 | OPMiniCPM5-1B | OpenBMB | 23.14 |
Evidence key: ObservedLast good
Rows are ordered by the value SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.