# Science Question Answering on ARC Hard

Source: [Science Question Answering on ARC Hard](<https://www.sota2.com/research/sota/science-question-answering-on-arc-hard>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-27T10:47:29.915Z

Evaluation of model performance on the difficult subset of the Abstract Reasoning Corpus (ARC Hard) for Science Question Answering.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Science Question Answering |
| Dataset (including any stated split) | ARC Hard |
| Metric | Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 4 of 4 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Score](<https://www.sota2.com/research/sota/science-question-answering-on-arc-hard.md?metric=Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Score | Result date | Report paper |
| --- | --- | --- | --- |
| BYOL-mri | 32.59 | Not recorded | 696dba82e25ad2fa9d1eccb7 |
| BYOL-mri | 32.17 | Not recorded | 696dba82e25ad2fa9d1eccb7 |
| BYOL-mri | 22.87 | Not recorded | 696dba82e25ad2fa9d1eccb7 |
| Llama-3.2 | 18.86 | Not recorded | 696dba82e25ad2fa9d1eccb7 |

### BYOL-mri

Score: 32.59. Result ID: 69a1763c35dfcabb00e731b6.

[Reporting paper](<https://www.sota2.com/research/paper/696dba82e25ad2fa9d1eccb7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/696dba82e25ad2fa9d1eccb7.md>)

Recorded evaluation configuration: {"Parameters":"12B","Training Stage":"CPT"}. This record does not assert independent reproduction.

### BYOL-mri

Score: 32.17. Result ID: 69a1763c35dfcabb00e731b5.

[Reporting paper](<https://www.sota2.com/research/paper/696dba82e25ad2fa9d1eccb7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/696dba82e25ad2fa9d1eccb7.md>)

Recorded evaluation configuration: {"Parameters":"4B","Training Stage":"CPT"}. This record does not assert independent reproduction.

### BYOL-mri

Score: 22.87. Result ID: 69a1763c35dfcabb00e731b4.

[Reporting paper](<https://www.sota2.com/research/paper/696dba82e25ad2fa9d1eccb7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/696dba82e25ad2fa9d1eccb7.md>)

Recorded evaluation configuration: {"Parameters":"1B","Training Stage":"CPT"}. This record does not assert independent reproduction.

### Llama-3.2

Score: 18.86. Result ID: 69a1763c35dfcabb00e731b3.

[Reporting paper](<https://www.sota2.com/research/paper/696dba82e25ad2fa9d1eccb7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75d11e209131cea6a028.md>)

[Implementation](<https://github.com/meta-llama/llama-agentic-system>)

Recorded evaluation configuration: {"Parameters":"1B"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.