# Instance Image Navigation on InsImageNav

Source: [Instance Image Navigation on InsImageNav](<https://www.sota2.com/research/sota/instance-image-navigation-on-insimagenav>)

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

Data updated: 2026-03-12T08:17:32.653Z

Evaluation of Instance Image Navigation performance on the InsImageNav dataset, measured by SR and SPL.

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 | Instance Image Navigation |
| Dataset (including any stated split) | InsImageNav |
| Metric | Success Rate (SR) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 9 of 9 |
| 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

- [Success Rate (SR)](<https://www.sota2.com/research/sota/instance-image-navigation-on-insimagenav.md?metric=Success%20Rate%20(SR)>): Higher is better; unit: PERCENT.

- [Success weighted Path Length (SPL)](<https://www.sota2.com/research/sota/instance-image-navigation-on-insimagenav.md?metric=Success%20weighted%20Path%20Length%20(SPL)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Success Rate (SR) | Result date | Report paper |
| --- | --- | --- | --- |
| CompassNav (SFT+RFT) | 35.6 | Not recorded | 68fa7c091e209131cea8bfe0 |
| GPT-o4-mini | 33.4 | Not recorded | 68fa7c091e209131cea8bfe0 |
| GPT-4o | 29.8 | Not recorded | 68fa7c091e209131cea8bfe0 |
| Gemini-2.5-Flash | 24.1 | Not recorded | 68fa7c091e209131cea8bfe0 |
| CompassNav (SFT) | 23.9 | Not recorded | 68fa7c091e209131cea8bfe0 |
| Base Model | 19.8 | Not recorded | 68fa7c091e209131cea8bfe0 |
| Qwen2.5-VL-3B | 8.7 | Not recorded | 68fa7c091e209131cea8bfe0 |
| Qwen2-VL-7B | 5.3 | Not recorded | 68fa7c091e209131cea8bfe0 |
| LLama3.2-11B | 3 | Not recorded | 68fa7c091e209131cea8bfe0 |

### CompassNav (SFT+RFT)

Success Rate (SR): 35.6. Result ID: 69b276856dbb7075b95cfa6d.

[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 Source":"Our Models","Training Protocol":"SFT+RFT","Backbone":"Qwen2.5-VL-7B"}. This record does not assert independent reproduction.

### GPT-o4-mini

Success Rate (SR): 33.4. Result ID: 69b276856dbb7075b95cfa6a.

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

Recorded evaluation configuration: {"Model Source":"Closed-Source Models","Model Scale":"mini"}. This record does not assert independent reproduction.

### GPT-4o

Success Rate (SR): 29.8. Result ID: 69b276856dbb7075b95cfa68.

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

[Method origin paper](<https://www.sota2.com/research/paper/68fa75681e209131cea67a98.md>)

Recorded evaluation configuration: {"Model Source":"Closed-Source Models"}. This record does not assert independent reproduction.

### Gemini-2.5-Flash

Success Rate (SR): 24.1. Result ID: 69b276856dbb7075b95cfa69.

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

[Method origin paper](<https://www.sota2.com/research/paper/68fa7c311e209131cea8cef5.md>)

Recorded evaluation configuration: {"Model Source":"Closed-Source Models"}. This record does not assert independent reproduction.

### CompassNav (SFT)

Success Rate (SR): 23.9. Result ID: 69b276856dbb7075b95cfa6c.

[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 Source":"Our Models","Training Protocol":"SFT","Backbone":"Qwen2.5-VL-7B"}. This record does not assert independent reproduction.

### Base Model

Success Rate (SR): 19.8. Result ID: 69b276856dbb7075b95cfa6b.

[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 Source":"Our Models","Backbone":"Qwen2.5-VL-7B"}. This record does not assert independent reproduction.

### Qwen2.5-VL-3B

Success Rate (SR): 8.7. Result ID: 69b276856dbb7075b95cfa66.

[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 Source":"Open-source Models","Model Scale":"3B"}. This record does not assert independent reproduction.

### Qwen2-VL-7B

Success Rate (SR): 5.3. Result ID: 69b276856dbb7075b95cfa65.

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

[Method origin paper](<https://www.sota2.com/research/paper/68fa75151e209131cea65976.md>)

[Implementation](<https://github.com/QwenLM/Qwen2-VL>)

Recorded evaluation configuration: {"Model Source":"Open-source Models","Model Scale":"7B"}. This record does not assert independent reproduction.

### LLama3.2-11B

Success Rate (SR): 3. Result ID: 69b276856dbb7075b95cfa67.

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

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

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

Recorded evaluation configuration: {"Model Source":"Open-source Models","Model Scale":"11B"}. This record does not assert independent reproduction.

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