# Challenging Reasoning Tasks on BBH (Accuracy and Loss)

Source: [Challenging Reasoning Tasks on BBH (Accuracy and Loss)](<https://www.sota2.com/research/sota/challenging-reasoning-tasks-on-bbh-accuracy-and-loss>)

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

Data updated: 2026-04-13T09:20:09.643Z

Evaluation of model performance on the Big-Bench Hard (BBH) challenging reasoning tasks, reporting Accuracy and Loss metrics.

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 | Challenging Reasoning Tasks |
| Dataset (including any stated split) | BBH |
| Metric | Accuracy |
| 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

- [Accuracy](<https://www.sota2.com/research/sota/challenging-reasoning-tasks-on-bbh-accuracy-and-loss.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [Loss](<https://www.sota2.com/research/sota/challenging-reasoning-tasks-on-bbh-accuracy-and-loss.md?metric=Loss>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| AdamW | 36.6 | Not recorded | 69dc78727d0172140ec78a09 |
| Nexus | 36.6 | Not recorded | 69dc78727d0172140ec78a09 |
| AdamW | 29.6 | Not recorded | 69dc78727d0172140ec78a09 |
| Nexus | 29.3 | Not recorded | 69dc78727d0172140ec78a09 |

### AdamW

Accuracy: 36.6. Result ID: 69dcb51b47f34ce924c3ebf9.

[Reporting paper](<https://www.sota2.com/research/paper/69dc78727d0172140ec78a09.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa55761e209131ce9946db.md>)

Recorded evaluation configuration: {"Model Size":"3B","Optimizer":"AdamW"}. This record does not assert independent reproduction.

### Nexus

Accuracy: 36.6. Result ID: 69dcb51b47f34ce924c3ebfa.

[Reporting paper](<https://www.sota2.com/research/paper/69dc78727d0172140ec78a09.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69dc78727d0172140ec78a09.md>)

Recorded evaluation configuration: {"Model Size":"3B","Optimizer":"Nexus"}. This record does not assert independent reproduction.

### AdamW

Accuracy: 29.6. Result ID: 69dcb51b47f34ce924c3ebf7.

[Reporting paper](<https://www.sota2.com/research/paper/69dc78727d0172140ec78a09.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa55761e209131ce9946db.md>)

Recorded evaluation configuration: {"Model Size":"1B","Optimizer":"AdamW"}. This record does not assert independent reproduction.

### Nexus

Accuracy: 29.3. Result ID: 69dcb51b47f34ce924c3ebf8.

[Reporting paper](<https://www.sota2.com/research/paper/69dc78727d0172140ec78a09.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69dc78727d0172140ec78a09.md>)

Recorded evaluation configuration: {"Model Size":"1B","Optimizer":"Nexus"}. This record does not assert independent reproduction.

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