# GUI Grounding on OSWorld G-Refine v1.0 (test)

Source: [GUI Grounding on OSWorld G-Refine v1.0 (test)](<https://www.sota2.com/research/sota/gui-grounding-on-osworld-g-refine-v1-0-test>)

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

Data updated: 2026-06-15T07:56:57.260Z

Evaluation of GUI Grounding performance on the test split of the OSWorld G-Refine v1.0 dataset, reporting component metrics including Text Matching, Element Recognition, Layout Understanding, Fine-grained Manipulation, and Refusal rates.

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 Grounding |
| Dataset (including any stated split) | OSWorld G-Refine v1.0 (test) |
| Metric | Overall Success Rate |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 37 of 37 |
| 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

- [Overall Success Rate](<https://www.sota2.com/research/sota/gui-grounding-on-osworld-g-refine-v1-0-test.md?metric=Overall%20Success%20Rate>): Higher is better; unit: PERCENT.

- [Text Matching](<https://www.sota2.com/research/sota/gui-grounding-on-osworld-g-refine-v1-0-test.md?metric=Text%20Matching>): Higher is better; unit: PERCENT.

- [Element Recognition](<https://www.sota2.com/research/sota/gui-grounding-on-osworld-g-refine-v1-0-test.md?metric=Element%20Recognition>): Higher is better; unit: PERCENT.

- [Layout Understanding](<https://www.sota2.com/research/sota/gui-grounding-on-osworld-g-refine-v1-0-test.md?metric=Layout%20Understanding>): Higher is better; unit: PERCENT.

- [Fine-grained Manipulation](<https://www.sota2.com/research/sota/gui-grounding-on-osworld-g-refine-v1-0-test.md?metric=Fine-grained%20Manipulation>): Higher is better; unit: PERCENT.

- [Refusal Rate](<https://www.sota2.com/research/sota/gui-grounding-on-osworld-g-refine-v1-0-test.md?metric=Refusal%20Rate>): Lower is better; unit: PERCENT.

## Recorded results

| Method | Overall Success Rate | Result date | Report paper |
| --- | --- | --- | --- |
| Holo2 | 76.1 | Not recorded | 6a2f8738d33dfb23e783c15f |
| MAI-UI-32B | 75 | Not recorded | 69520ae053fdd5c36173d429 |
| MAI-UI-32B | 73.9 | Not recorded | 69520ae053fdd5c36173d429 |
| MAI-UI-8B | 72.9 | Not recorded | 69520ae053fdd5c36173d429 |
| VISTA | 72.5 | Not recorded | 6a2f8738d33dfb23e783c15f |
| GTA1-32B | 72.2 | Not recorded | 69520ae053fdd5c36173d429 |
| GTA1 | 72.2 | Not recorded | 6a2f8738d33dfb23e783c15f |
| VISTA | 72 | Not recorded | 6a2f8738d33dfb23e783c15f |
| VISTA | 70.9 | Not recorded | 6a2f8738d33dfb23e783c15f |
| VISTA | 70.8 | Not recorded | 6a2f8738d33dfb23e783c15f |
| OpenCUA-32B | 70.2 | Not recorded | 69520ae053fdd5c36173d429 |
| OpenCUA | 70.2 | Not recorded | 6a2f8738d33dfb23e783c15f |
| Holo2 | 70.1 | Not recorded | 6a2f8738d33dfb23e783c15f |
| Step-GUI | 70 | Not recorded | 6a2f8738d33dfb23e783c15f |
| VISTA | 69.6 | Not recorded | 6a2f8738d33dfb23e783c15f |
| UI-Venus | 69.5 | Not recorded | 6a2f8738d33dfb23e783c15f |
| VISTA | 69.4 | Not recorded | 6a2f8738d33dfb23e783c15f |
| Qwen3-VL | 69.3 | Not recorded | 6a2f8738d33dfb23e783c15f |
| Qwen3-VL-32B | 69 | Not recorded | 69520ae053fdd5c36173d429 |
| MAI-UI-8B | 68.6 | Not recorded | 69520ae053fdd5c36173d429 |
| MAI-UI | 68.6 | Not recorded | 6a2f8738d33dfb23e783c15f |
| Qwen3-VL | 67.9 | Not recorded | 6a2f8738d33dfb23e783c15f |
| GTA1-7B | 67.7 | Not recorded | 69520ae053fdd5c36173d429 |
| GTA1 | 67.7 | Not recorded | 6a2f8738d33dfb23e783c15f |
| Step-GUI | 66.9 | Not recorded | 6a2f8738d33dfb23e783c15f |
| MAI-UI-2B | 66.3 | Not recorded | 69520ae053fdd5c36173d429 |
| Qwen3-VL-8B | 64.4 | Not recorded | 69520ae053fdd5c36173d429 |
| Qwen3-VL | 64.4 | Not recorded | 6a2f8738d33dfb23e783c15f |
| UI-TARS-1.5-7B | 64.2 | Not recorded | 69520ae053fdd5c36173d429 |
| UI-TARS-1.5 | 64.2 | Not recorded | 6a2f8738d33dfb23e783c15f |
| Jedi-7B | 63.8 | Not recorded | 69520ae053fdd5c36173d429 |
| MAI-UI-2B | 63.5 | Not recorded | 69520ae053fdd5c36173d429 |
| UI-Venus | 61.7 | Not recorded | 6a2f8738d33dfb23e783c15f |
| Jedi-3B | 61 | Not recorded | 69520ae053fdd5c36173d429 |
| Qwen2.5-VL-32B | 59.6 | Not recorded | 69520ae053fdd5c36173d429 |
| Operator | 57.8 | Not recorded | 69520ae053fdd5c36173d429 |
| Qwen3-VL-2B | 57.4 | Not recorded | 69520ae053fdd5c36173d429 |

### Holo2

Overall Success Rate: 76.1. Result ID: 6a2fb02dfbc373592c2bd24c.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

Recorded evaluation configuration: {"Model Scale":"≥30B"}. This record does not assert independent reproduction.

### MAI-UI-32B

Overall Success Rate: 75. Result ID: 6998094d5698665ec462dba0.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

Recorded evaluation configuration: {"Zoom-In":true}. This record does not assert independent reproduction.

### MAI-UI-32B

Overall Success Rate: 73.9. Result ID: 6998094d5698665ec462db9f.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

Recorded evaluation configuration: {"Zoom-In":false}. This record does not assert independent reproduction.

### MAI-UI-8B

Overall Success Rate: 72.9. Result ID: 6998094d5698665ec462db9e.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

Recorded evaluation configuration: {"Zoom-In":true}. This record does not assert independent reproduction.

### VISTA

Overall Success Rate: 72.5. Result ID: 6a2fb02dfbc373592c2bd251.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Implementation](<https://github.com/ZJUSCL/VISTA>)

Recorded evaluation configuration: {"Model Scale":"≥30B","Inference Method":"Multi-view (+ MVP)"}. This record does not assert independent reproduction.

### GTA1-32B

Overall Success Rate: 72.2. Result ID: 6998094d5698665ec462db9a.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

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

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

### GTA1

Overall Success Rate: 72.2. Result ID: 6a2fb02dfbc373592c2bd24e.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

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

Recorded evaluation configuration: {"Model Scale":"≥30B"}. This record does not assert independent reproduction.

### VISTA

Overall Success Rate: 72. Result ID: 6a2fb02dfbc373592c2bd250.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Implementation](<https://github.com/ZJUSCL/VISTA>)

Recorded evaluation configuration: {"Model Scale":"≥30B","Inference Method":"Single-view"}. This record does not assert independent reproduction.

### VISTA

Overall Success Rate: 70.9. Result ID: 6a2fb02dfbc373592c2bd24a.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Implementation](<https://github.com/ZJUSCL/VISTA>)

Recorded evaluation configuration: {"Model Scale":"≈8B","Inference Method":"Multi-view (+ MVP)"}. This record does not assert independent reproduction.

### VISTA

Overall Success Rate: 70.8. Result ID: 6a2fb02dfbc373592c2bd249.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Implementation](<https://github.com/ZJUSCL/VISTA>)

Recorded evaluation configuration: {"Model Scale":"≈8B","Inference Method":"Single-view"}. This record does not assert independent reproduction.

### OpenCUA-32B

Overall Success Rate: 70.2. Result ID: 6998094d5698665ec462db98.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

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

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

### OpenCUA

Overall Success Rate: 70.2. Result ID: 6a2fb02dfbc373592c2bd24d.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

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

Recorded evaluation configuration: {"Model Scale":"≥30B"}. This record does not assert independent reproduction.

### Holo2

Overall Success Rate: 70.1. Result ID: 6a2fb02dfbc373592c2bd247.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

Recorded evaluation configuration: {"Model Scale":"≈8B"}. This record does not assert independent reproduction.

### Step-GUI

Overall Success Rate: 70. Result ID: 6a2fb02dfbc373592c2bd245.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69438a9b423ea422657294f6.md>)

[Implementation](<https://github.com/stepfun-ai/gelab-zero>)

Recorded evaluation configuration: {"Model Scale":"≈8B"}. This record does not assert independent reproduction.

### VISTA

Overall Success Rate: 69.6. Result ID: 6a2fb02dfbc373592c2bd241.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Implementation](<https://github.com/ZJUSCL/VISTA>)

Recorded evaluation configuration: {"Model Scale":"≈4B","Inference Method":"Multi-view (+ MVP)"}. This record does not assert independent reproduction.

### UI-Venus

Overall Success Rate: 69.5. Result ID: 6a2fb02dfbc373592c2bd24f.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ab01e209131cea849f9.md>)

Recorded evaluation configuration: {"Model Scale":"72B"}. This record does not assert independent reproduction.

### VISTA

Overall Success Rate: 69.4. Result ID: 6a2fb02dfbc373592c2bd240.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Implementation](<https://github.com/ZJUSCL/VISTA>)

Recorded evaluation configuration: {"Model Scale":"≈4B","Inference Method":"Single-view"}. This record does not assert independent reproduction.

### Qwen3-VL

Overall Success Rate: 69.3. Result ID: 6a2fb02dfbc373592c2bd24b.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"Model Scale":"≥30B"}. This record does not assert independent reproduction.

### Qwen3-VL-32B

Overall Success Rate: 69. Result ID: 6998094d5698665ec462db99.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"Evaluated by authors":true}. This record does not assert independent reproduction.

### MAI-UI-8B

Overall Success Rate: 68.6. Result ID: 6998094d5698665ec462db9d.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

Recorded evaluation configuration: {"Zoom-In":false}. This record does not assert independent reproduction.

### MAI-UI

Overall Success Rate: 68.6. Result ID: 6a2fb02dfbc373592c2bd248.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

Recorded evaluation configuration: {"Model Scale":"≈8B"}. This record does not assert independent reproduction.

### Qwen3-VL

Overall Success Rate: 67.9. Result ID: 6a2fb02dfbc373592c2bd23e.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"Model Scale":"≈4B"}. This record does not assert independent reproduction.

### GTA1-7B

Overall Success Rate: 67.7. Result ID: 6998094d5698665ec462db96.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

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

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

### GTA1

Overall Success Rate: 67.7. Result ID: 6a2fb02dfbc373592c2bd243.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

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

Recorded evaluation configuration: {"Model Scale":"≈8B"}. This record does not assert independent reproduction.

### Step-GUI

Overall Success Rate: 66.9. Result ID: 6a2fb02dfbc373592c2bd23f.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69438a9b423ea422657294f6.md>)

[Implementation](<https://github.com/stepfun-ai/gelab-zero>)

Recorded evaluation configuration: {"Model Scale":"≈4B"}. This record does not assert independent reproduction.

### MAI-UI-2B

Overall Success Rate: 66.3. Result ID: 6998094d5698665ec462db9c.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

Recorded evaluation configuration: {"Zoom-In":true}. This record does not assert independent reproduction.

### Qwen3-VL-8B

Overall Success Rate: 64.4. Result ID: 6998094d5698665ec462db95.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"Evaluated by authors":true}. This record does not assert independent reproduction.

### Qwen3-VL

Overall Success Rate: 64.4. Result ID: 6a2fb02dfbc373592c2bd246.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"Model Scale":"≈8B"}. This record does not assert independent reproduction.

### UI-TARS-1.5-7B

Overall Success Rate: 64.2. Result ID: 6998094d5698665ec462db94.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

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

### UI-TARS-1.5

Overall Success Rate: 64.2. Result ID: 6a2fb02dfbc373592c2bd242.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76a61e209131cea6e2ba.md>)

Recorded evaluation configuration: {"Model Scale":"≈8B"}. This record does not assert independent reproduction.

### Jedi-7B

Overall Success Rate: 63.8. Result ID: 6998094d5698665ec462db93.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79861e209131cea7e01e.md>)

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

### MAI-UI-2B

Overall Success Rate: 63.5. Result ID: 6998094d5698665ec462db9b.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

Recorded evaluation configuration: {"Zoom-In":false}. This record does not assert independent reproduction.

### UI-Venus

Overall Success Rate: 61.7. Result ID: 6a2fb02dfbc373592c2bd244.

[Reporting paper](<https://www.sota2.com/research/paper/6a2f8738d33dfb23e783c15f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ab01e209131cea849f9.md>)

Recorded evaluation configuration: {"Model Scale":"≈8B"}. This record does not assert independent reproduction.

### Jedi-3B

Overall Success Rate: 61. Result ID: 6998094d5698665ec462db92.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79861e209131cea7e01e.md>)

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

### Qwen2.5-VL-32B

Overall Success Rate: 59.6. Result ID: 6998094d5698665ec462db97.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

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

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

### Operator

Overall Success Rate: 57.8. Result ID: 6998094d5698665ec462db90.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

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

### Qwen3-VL-2B

Overall Success Rate: 57.4. Result ID: 6998094d5698665ec462db91.

[Reporting paper](<https://www.sota2.com/research/paper/69520ae053fdd5c36173d429.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"Evaluated by authors":true}. This record does not assert independent reproduction.

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