# Decision Understanding on NavNuances

Source: [Decision Understanding on NavNuances](<https://www.sota2.com/research/sota/decision-understanding-on-navnuances>)

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

Data updated: 2026-03-12T08:23:52.254Z

Evaluation of Decision Understanding capabilities on the NavNuances dataset, reporting metrics such as DC, NU, LR, RR, and VM.

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 | Decision Understanding |
| Dataset (including any stated split) | NavNuances |
| Metric | DC |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| 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

- [DC](<https://www.sota2.com/research/sota/decision-understanding-on-navnuances.md?metric=DC>): Higher is better; unit: PERCENT.

- [NU](<https://www.sota2.com/research/sota/decision-understanding-on-navnuances.md?metric=NU>): Higher is better; unit: PERCENT.

- [LR](<https://www.sota2.com/research/sota/decision-understanding-on-navnuances.md?metric=LR>): Higher is better; unit: PERCENT.

- [RR](<https://www.sota2.com/research/sota/decision-understanding-on-navnuances.md?metric=RR>): Higher is better; unit: PERCENT.

- [VM](<https://www.sota2.com/research/sota/decision-understanding-on-navnuances.md?metric=VM>): Lower is better; unit: PERCENT.

## Recorded results

| Method | DC | Result date | Report paper |
| --- | --- | --- | --- |
| NavGPT-4V | 92.68 | Not recorded | 68fa7c091e209131cea8bfe0 |
| NavGPT-4 | 91.87 | Not recorded | 68fa7c091e209131cea8bfe0 |
| CompassNav-7B | 82.54 | Not recorded | 68fa7c091e209131cea8bfe0 |
| NavGPT-3.5 | 81.87 | Not recorded | 68fa7c091e209131cea8bfe0 |
| Qwen2.5-VL-7B | 79.62 | Not recorded | 68fa7c091e209131cea8bfe0 |

### NavGPT-4V

DC: 92.68. Result ID: 69b276856dbb7075b95cfa79.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c091e209131cea8bfe0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74fd1e209131cea651b3.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### NavGPT-4

DC: 91.87. Result ID: 69b276856dbb7075b95cfa78.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c091e209131cea8bfe0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74fd1e209131cea651b3.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### CompassNav-7B

DC: 82.54. Result ID: 69b276856dbb7075b95cfa7b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c091e209131cea8bfe0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c091e209131cea8bfe0.md>)

Recorded evaluation configuration: {"Model Size":"7B"}. This record does not assert independent reproduction.

### NavGPT-3.5

DC: 81.87. Result ID: 69b276856dbb7075b95cfa77.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c091e209131cea8bfe0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74fd1e209131cea651b3.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Qwen2.5-VL-7B

DC: 79.62. Result ID: 69b276856dbb7075b95cfa7a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c091e209131cea8bfe0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa773a1e209131cea71d3c.md>)

Recorded evaluation configuration: {"Model Size":"7B"}. This record does not assert independent reproduction.

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