# Long-context reasoning on RULER 8K context length

Source: [Long-context reasoning on RULER 8K context length](<https://www.sota2.com/research/sota/long-context-reasoning-on-ruler-8k-context-length>)

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

Data updated: 2026-05-20T09:14:13.011Z

Evaluation of model performance on the RULER dataset specifically configured with an 8,000 token context length, focusing on long-context reasoning capabilities, measured by NIAH, AGG, QA, VT, and AVG scores.

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 | Long-context reasoning |
| Dataset (including any stated split) | RULER 8K context length |
| Metric | NIAH Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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

- [NIAH Score](<https://www.sota2.com/research/sota/long-context-reasoning-on-ruler-8k-context-length.md?metric=NIAH%20Score>): Higher is better; unit: PERCENT.

- [AGG Score](<https://www.sota2.com/research/sota/long-context-reasoning-on-ruler-8k-context-length.md?metric=AGG%20Score>): Higher is better; unit: PERCENT.

- [QA Score](<https://www.sota2.com/research/sota/long-context-reasoning-on-ruler-8k-context-length.md?metric=QA%20Score>): Higher is better; unit: PERCENT.

- [VT Score](<https://www.sota2.com/research/sota/long-context-reasoning-on-ruler-8k-context-length.md?metric=VT%20Score>): Higher is better; unit: PERCENT.

- [AVG Score](<https://www.sota2.com/research/sota/long-context-reasoning-on-ruler-8k-context-length.md?metric=AVG%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | NIAH Score | Result date | Report paper |
| --- | --- | --- | --- |
| UltraLLaDA+BA-Att | 98.5 | Not recorded | 6a0d4054f984739ed6737b45 |
| UltraLLaDA | 98.28 | Not recorded | 6a0d4054f984739ed6737b45 |
| UltraLLaDA+XAtt | 86.97 | Not recorded | 6a0d4054f984739ed6737b45 |
| LLaDA1.5 | 51.56 | Not recorded | 6a0d4054f984739ed6737b45 |
| LLaDA1.5+BA-Att | 50.28 | Not recorded | 6a0d4054f984739ed6737b45 |
| LLaDA1.5+XAtt | 39.53 | Not recorded | 6a0d4054f984739ed6737b45 |

### UltraLLaDA+BA-Att

NIAH Score: 98.5. Result ID: 6a0d7a04fbc373592c05b889.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4054f984739ed6737b45.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0d4054f984739ed6737b45.md>)

Recorded evaluation configuration: {"Base Model":"UltraLLaDA","Attention Method":"BA-Att"}. This record does not assert independent reproduction.

### UltraLLaDA

NIAH Score: 98.28. Result ID: 6a0d7a04fbc373592c05b887.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4054f984739ed6737b45.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c0a1e209131cea8c08b.md>)

Recorded evaluation configuration: {"Base Model":"UltraLLaDA","Attention Method":"Full Attention"}. This record does not assert independent reproduction.

### UltraLLaDA+XAtt

NIAH Score: 86.97. Result ID: 6a0d7a04fbc373592c05b888.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4054f984739ed6737b45.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77811e209131cea73a62.md>)

Recorded evaluation configuration: {"Base Model":"UltraLLaDA","Attention Method":"XAtt"}. This record does not assert independent reproduction.

### LLaDA1.5

NIAH Score: 51.56. Result ID: 6a0d7a04fbc373592c05b884.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4054f984739ed6737b45.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c0e1e209131cea8c28e.md>)

Recorded evaluation configuration: {"Base Model":"LLaDA1.5","Attention Method":"Full Attention"}. This record does not assert independent reproduction.

### LLaDA1.5+BA-Att

NIAH Score: 50.28. Result ID: 6a0d7a04fbc373592c05b886.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4054f984739ed6737b45.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0d4054f984739ed6737b45.md>)

Recorded evaluation configuration: {"Base Model":"LLaDA1.5","Attention Method":"BA-Att"}. This record does not assert independent reproduction.

### LLaDA1.5+XAtt

NIAH Score: 39.53. Result ID: 6a0d7a04fbc373592c05b885.

[Reporting paper](<https://www.sota2.com/research/paper/6a0d4054f984739ed6737b45.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77811e209131cea73a62.md>)

Recorded evaluation configuration: {"Base Model":"LLaDA1.5","Attention Method":"XAtt"}. This record does not assert independent reproduction.

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