# GUI Navigation on OSWorld Verified

Source: [GUI Navigation on OSWorld Verified](<https://www.sota2.com/research/sota/gui-navigation-on-osworld-verified>)

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

Data updated: 2026-04-24T08:08:07.452Z

Evaluation of agent performance on GUI Navigation tasks using the OSWorld Verified dataset, reporting success rates across various operational domains (OS, Office, Daily, Professional, Work).

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 | GUI Navigation |
| Dataset (including any stated split) | OSWorld Verified |
| Metric | OS Success Rate |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 32 of 32 |
| 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

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

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

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

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

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

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

## Recorded results

| Method | OS Success Rate | Result date | Report paper |
| --- | --- | --- | --- |
| VLAA-GUI w/ Gemini 3 Flash | 91.7 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Opus 4.6 | 91.7 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Opus 4.5 + MAI-UI | 91.67 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Opus 4.5 | 87.5 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Opus 4.5 | 87.5 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| HIPPO w/ Opus 4.5 | 87.5 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Opus 4.5 | 87.5 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Gemini 3.1 Pro | 83.3 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Sonnet 4.6 | 83.3 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Opus 4.6 | 83.3 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Gemini 3 Flash | 83.3 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Gemini 3.1 Pro | 83.3 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Sonnet 4.6 | 83.3 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Opus 4.6 | 83.3 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Gemini 3.1 Pro | 83.3 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Sonnet 4.6 | 83.3 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| VLAA-GUI w/ Gemini 3 Flash | 79.2 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| GTA1 w/ GPT-5 | 79.17 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| OS-Symphony w/ GPT-5 | 79.17 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| EvoCUA | 78.26 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| Agent S3 w/ GPT-5 | 77.5 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| OS-Symphony w/ GPT-5 | 75 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| CoAct-1 w/ GPT-5 | 75 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| Agent S3 w/ Opus 4.5 | 75 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| UiPath w/ GPT-5 | 73.91 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| Kimi K2.5 | 73.91 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| CoACT-1 w/ GPT-5 | 70.83 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| Claude-Sonnet-4.5 | 70.83 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| UiPath w/ Opus 4.5 | 70.83 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| Seed-1.8 | 66.67 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| DeepMiner-Mano-72B | 66.67 | Not recorded | 69eaf92c55a3c32af89f8a00 |
| OpenCUA-72B | 61.13 | Not recorded | 69eaf92c55a3c32af89f8a00 |

### VLAA-GUI w/ Gemini 3 Flash

OS Success Rate: 91.7. Result ID: 69eb24d647f34ce924d33678.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Opus 4.6

OS Success Rate: 91.7. Result ID: 69eb24d647f34ce924d3367d.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Opus 4.5 + MAI-UI

OS Success Rate: 91.67. Result ID: 69eb24d647f34ce924d3367c.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Opus 4.5

OS Success Rate: 87.5. Result ID: 69eb24d647f34ce924d33661.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"15"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Opus 4.5

OS Success Rate: 87.5. Result ID: 69eb24d647f34ce924d3366a.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"50"}. This record does not assert independent reproduction.

### HIPPO w/ Opus 4.5

OS Success Rate: 87.5. Result ID: 69eb24d647f34ce924d33677.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Opus 4.5

OS Success Rate: 87.5. Result ID: 69eb24d647f34ce924d3367b.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Gemini 3.1 Pro

OS Success Rate: 83.3. Result ID: 69eb24d647f34ce924d3365f.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"15"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Sonnet 4.6

OS Success Rate: 83.3. Result ID: 69eb24d647f34ce924d33660.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"15"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Opus 4.6

OS Success Rate: 83.3. Result ID: 69eb24d647f34ce924d33662.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"15"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Gemini 3 Flash

OS Success Rate: 83.3. Result ID: 69eb24d647f34ce924d33667.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"50"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Gemini 3.1 Pro

OS Success Rate: 83.3. Result ID: 69eb24d647f34ce924d33668.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"50"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Sonnet 4.6

OS Success Rate: 83.3. Result ID: 69eb24d647f34ce924d33669.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"50"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Opus 4.6

OS Success Rate: 83.3. Result ID: 69eb24d647f34ce924d3366b.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"50"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Gemini 3.1 Pro

OS Success Rate: 83.3. Result ID: 69eb24d647f34ce924d33679.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Sonnet 4.6

OS Success Rate: 83.3. Result ID: 69eb24d647f34ce924d3367a.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### VLAA-GUI w/ Gemini 3 Flash

OS Success Rate: 79.2. Result ID: 69eb24d647f34ce924d3365e.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Implementation](<https://github.com/UCSC-VLAA/VLAA-GUI>)

Recorded evaluation configuration: {"Step Budget":"15"}. This record does not assert independent reproduction.

### GTA1 w/ GPT-5

OS Success Rate: 79.17. Result ID: 69eb24d647f34ce924d33672.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7be61e209131cea8b2ba.md>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### OS-Symphony w/ GPT-5

OS Success Rate: 79.17. Result ID: 69eb24d647f34ce924d33673.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6965d1b56343fa78c8323964.md>)

[Implementation](<https://github.com/OS-Copilot/OS-Symphony>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### EvoCUA

OS Success Rate: 78.26. Result ID: 69eb24d647f34ce924d33663.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6973007bb2be0ed85a78c9c4.md>)

[Implementation](<https://github.com/meituan/EvoCUA>)

Recorded evaluation configuration: {"Step Budget":"50"}. This record does not assert independent reproduction.

### Agent S3 w/ GPT-5

OS Success Rate: 77.5. Result ID: 69eb24d647f34ce924d33674.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bd61e209131cea8ad11.md>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### OS-Symphony w/ GPT-5

OS Success Rate: 75. Result ID: 69eb24d647f34ce924d33666.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6965d1b56343fa78c8323964.md>)

[Implementation](<https://github.com/OS-Copilot/OS-Symphony>)

Recorded evaluation configuration: {"Step Budget":"50"}. This record does not assert independent reproduction.

### CoAct-1 w/ GPT-5

OS Success Rate: 75. Result ID: 69eb24d647f34ce924d33671.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### Agent S3 w/ Opus 4.5

OS Success Rate: 75. Result ID: 69eb24d647f34ce924d33676.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bd61e209131cea8ad11.md>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### UiPath w/ GPT-5

OS Success Rate: 73.91. Result ID: 69eb24d647f34ce924d33664.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

Recorded evaluation configuration: {"Step Budget":"50"}. This record does not assert independent reproduction.

### Kimi K2.5

OS Success Rate: 73.91. Result ID: 69eb24d647f34ce924d33670.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/698181b4a8b01e76cc8e2812.md>)

[Implementation](<https://github.com/MoonshotAI/Kimi-K2.5>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### CoACT-1 w/ GPT-5

OS Success Rate: 70.83. Result ID: 69eb24d647f34ce924d33665.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

Recorded evaluation configuration: {"Step Budget":"50"}. This record does not assert independent reproduction.

### Claude-Sonnet-4.5

OS Success Rate: 70.83. Result ID: 69eb24d647f34ce924d3366f.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### UiPath w/ Opus 4.5

OS Success Rate: 70.83. Result ID: 69eb24d647f34ce924d33675.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### Seed-1.8

OS Success Rate: 66.67. Result ID: 69eb24d647f34ce924d3366d.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69c21a620636e5bf4afc461d.md>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### DeepMiner-Mano-72B

OS Success Rate: 66.67. Result ID: 69eb24d647f34ce924d3366e.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b871e209131cea88e94.md>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

### OpenCUA-72B

OS Success Rate: 61.13. Result ID: 69eb24d647f34ce924d3366c.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf92c55a3c32af89f8a00.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7be61e209131cea8b2f4.md>)

Recorded evaluation configuration: {"Step Budget":"100"}. This record does not assert independent reproduction.

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