# Accuracy on NQ (val)

Source: [Accuracy on NQ (val)](<https://www.sota2.com/research/sota/accuracy-on-nq-val>)

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

Data updated: 2026-04-21T07:55:58.705Z

Evaluation of model performance on the Natural Questions (NQ) validation set for Question Answering, reporting accuracy.

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 | Question Answering |
| Dataset (including any stated split) | NQ (val) |
| Metric | Accuracy (NQ val) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 18 of 18 |
| 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 (NQ val)](<https://www.sota2.com/research/sota/accuracy-on-nq-val.md?metric=Accuracy%20(NQ%20val)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy (NQ val) | Result date | Report paper |
| --- | --- | --- | --- |
| NWCAD | 79.95 | Not recorded | 69e704667d0172140ec7eb3c |
| NWCAD | 79.82 | Not recorded | 69e704667d0172140ec7eb3c |
| With-context | 79.29 | Not recorded | 69e704667d0172140ec7eb3c |
| With-context | 79.23 | Not recorded | 69e704667d0172140ec7eb3c |
| AdaCAD | 79.17 | Not recorded | 69e704667d0172140ec7eb3c |
| NWCAD | 78.88 | Not recorded | 69e704667d0172140ec7eb3c |
| With-context | 77.25 | Not recorded | 69e704667d0172140ec7eb3c |
| AdaCAD | 76.88 | Not recorded | 69e704667d0172140ec7eb3c |
| AdaCAD | 76.6 | Not recorded | 69e704667d0172140ec7eb3c |
| CoCoA | 73.82 | Not recorded | 69e704667d0172140ec7eb3c |
| CAD | 71.28 | Not recorded | 69e704667d0172140ec7eb3c |
| CoCoA | 69 | Not recorded | 69e704667d0172140ec7eb3c |
| CAD | 65.68 | Not recorded | 69e704667d0172140ec7eb3c |
| CoCoA | 63.6 | Not recorded | 69e704667d0172140ec7eb3c |
| CAD | 52.25 | Not recorded | 69e704667d0172140ec7eb3c |
| Baseline | 38.01 | Not recorded | 69e704667d0172140ec7eb3c |
| Baseline | 34.62 | Not recorded | 69e704667d0172140ec7eb3c |
| Baseline | 29.34 | Not recorded | 69e704667d0172140ec7eb3c |

### NWCAD

Accuracy (NQ val): 79.95. Result ID: 69e72d8c47f34ce924cd305b.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

Recorded evaluation configuration: {"Model":"Ministral-3-8B"}. This record does not assert independent reproduction.

### NWCAD

Accuracy (NQ val): 79.82. Result ID: 69e72d8c47f34ce924cd304f.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

Recorded evaluation configuration: {"Model":"Llama-3.1-8B"}. This record does not assert independent reproduction.

### With-context

Accuracy (NQ val): 79.29. Result ID: 69e72d8c47f34ce924cd3057.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

Recorded evaluation configuration: {"Model":"Ministral-3-8B"}. This record does not assert independent reproduction.

### With-context

Accuracy (NQ val): 79.23. Result ID: 69e72d8c47f34ce924cd304b.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

Recorded evaluation configuration: {"Model":"Llama-3.1-8B"}. This record does not assert independent reproduction.

### AdaCAD

Accuracy (NQ val): 79.17. Result ID: 69e72d8c47f34ce924cd3059.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78411e209131cea776f7.md>)

Recorded evaluation configuration: {"Model":"Ministral-3-8B"}. This record does not assert independent reproduction.

### NWCAD

Accuracy (NQ val): 78.88. Result ID: 69e72d8c47f34ce924cd3055.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

Recorded evaluation configuration: {"Model":"Llama-3.1-70B"}. This record does not assert independent reproduction.

### With-context

Accuracy (NQ val): 77.25. Result ID: 69e72d8c47f34ce924cd3051.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

Recorded evaluation configuration: {"Model":"Llama-3.1-70B"}. This record does not assert independent reproduction.

### AdaCAD

Accuracy (NQ val): 76.88. Result ID: 69e72d8c47f34ce924cd3053.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78411e209131cea776f7.md>)

Recorded evaluation configuration: {"Model":"Llama-3.1-70B"}. This record does not assert independent reproduction.

### AdaCAD

Accuracy (NQ val): 76.6. Result ID: 69e72d8c47f34ce924cd304d.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78411e209131cea776f7.md>)

Recorded evaluation configuration: {"Model":"Llama-3.1-8B"}. This record does not assert independent reproduction.

### CoCoA

Accuracy (NQ val): 73.82. Result ID: 69e72d8c47f34ce924cd304e.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7add1e209131cea8594e.md>)

Recorded evaluation configuration: {"Model":"Llama-3.1-8B"}. This record does not assert independent reproduction.

### CAD

Accuracy (NQ val): 71.28. Result ID: 69e72d8c47f34ce924cd3058.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f6e1e209131cea45813.md>)

Recorded evaluation configuration: {"Model":"Ministral-3-8B"}. This record does not assert independent reproduction.

### CoCoA

Accuracy (NQ val): 69. Result ID: 69e72d8c47f34ce924cd3054.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7add1e209131cea8594e.md>)

Recorded evaluation configuration: {"Model":"Llama-3.1-70B"}. This record does not assert independent reproduction.

### CAD

Accuracy (NQ val): 65.68. Result ID: 69e72d8c47f34ce924cd304c.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f6e1e209131cea45813.md>)

Recorded evaluation configuration: {"Model":"Llama-3.1-8B"}. This record does not assert independent reproduction.

### CoCoA

Accuracy (NQ val): 63.6. Result ID: 69e72d8c47f34ce924cd305a.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7add1e209131cea8594e.md>)

Recorded evaluation configuration: {"Model":"Ministral-3-8B"}. This record does not assert independent reproduction.

### CAD

Accuracy (NQ val): 52.25. Result ID: 69e72d8c47f34ce924cd3052.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f6e1e209131cea45813.md>)

Recorded evaluation configuration: {"Model":"Llama-3.1-70B"}. This record does not assert independent reproduction.

### Baseline

Accuracy (NQ val): 38.01. Result ID: 69e72d8c47f34ce924cd304a.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

Recorded evaluation configuration: {"Model":"Llama-3.1-8B"}. This record does not assert independent reproduction.

### Baseline

Accuracy (NQ val): 34.62. Result ID: 69e72d8c47f34ce924cd3050.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

Recorded evaluation configuration: {"Model":"Llama-3.1-70B"}. This record does not assert independent reproduction.

### Baseline

Accuracy (NQ val): 29.34. Result ID: 69e72d8c47f34ce924cd3056.

[Reporting paper](<https://www.sota2.com/research/paper/69e704667d0172140ec7eb3c.md>)

Recorded evaluation configuration: {"Model":"Ministral-3-8B"}. This record does not assert independent reproduction.

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