# Multimodal Tool Use on LLaVA-Bench Tool Use

Source: [Multimodal Tool Use on LLaVA-Bench Tool Use](<https://www.sota2.com/research/sota/multimodal-tool-use-on-llava-bench-tool-use>)

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

Data updated: 2026-02-26T05:46:04.179Z

Evaluation of multimodal model performance specifically on tool utilization capabilities using the LLaVA-Bench Tool Use dataset, measuring performance across grounding, tagging, captioning, and OCR tasks.

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 | Multimodal Tool Use |
| Dataset (including any stated split) | LLaVA-Bench Tool Use |
| Metric | Grounding |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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

- [Grounding](<https://www.sota2.com/research/sota/multimodal-tool-use-on-llava-bench-tool-use.md?metric=Grounding>): Higher is better; unit: PERCENT.

- [Tagging](<https://www.sota2.com/research/sota/multimodal-tool-use-on-llava-bench-tool-use.md?metric=Tagging>): Higher is better; unit: PERCENT.

- [Captioning Score](<https://www.sota2.com/research/sota/multimodal-tool-use-on-llava-bench-tool-use.md?metric=Captioning%20Score>): Higher is better; unit: PERCENT.

- [OCR](<https://www.sota2.com/research/sota/multimodal-tool-use-on-llava-bench-tool-use.md?metric=OCR>): Higher is better; unit: PERCENT.

- [Overall Performance](<https://www.sota2.com/research/sota/multimodal-tool-use-on-llava-bench-tool-use.md?metric=Overall%20Performance>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Grounding | Result date | Report paper |
| --- | --- | --- | --- |
| LLaVA-Plus | 89.3 | Not recorded | 68fa71201e209131cea4f4e0 |
| LLaVA-Plus | 88.6 | Not recorded | 68fa71201e209131cea4f4e0 |
| All Tools + GPT4 | 77.5 | Not recorded | 68fa71201e209131cea4f4e0 |
| Bing Chat | 56 | Not recorded | 68fa71201e209131cea4f4e0 |
| LLaVA | 47.1 | Not recorded | 68fa71201e209131cea4f4e0 |
| LLaVA | 41.7 | Not recorded | 68fa71201e209131cea4f4e0 |
| Bard | 36.5 | Not recorded | 68fa71201e209131cea4f4e0 |
| MM-REACT | 30.2 | Not recorded | 68fa71201e209131cea4f4e0 |

### LLaVA-Plus

Grounding: 89.3. Result ID: 6991b4c33a6c3148490a0d76.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71201e209131cea4f4e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71201e209131cea4f4e0.md>)

[Implementation](<https://llava-vl.github.io/llava-plus/>)

Recorded evaluation configuration: {"tool\_configuration":"All Tools"}. This record does not assert independent reproduction.

### LLaVA-Plus

Grounding: 88.6. Result ID: 6991b4c33a6c3148490a0d77.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71201e209131cea4f4e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71201e209131cea4f4e0.md>)

[Implementation](<https://llava-vl.github.io/llava-plus/>)

Recorded evaluation configuration: {"tool\_configuration":"Fly"}. This record does not assert independent reproduction.

### All Tools + GPT4

Grounding: 77.5. Result ID: 6991b4c33a6c3148490a0d7b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71201e209131cea4f4e0.md>)

Recorded evaluation configuration: {"planner\_configuration":"GPT-4","tool\_configuration":"All Tools"}. This record does not assert independent reproduction.

### Bing Chat

Grounding: 56. Result ID: 6991b4c33a6c3148490a0d79.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71201e209131cea4f4e0.md>)

Recorded evaluation configuration: {"evaluation\_date":"0730"}. This record does not assert independent reproduction.

### LLaVA

Grounding: 47.1. Result ID: 6991b4c33a6c3148490a0d74.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71201e209131cea4f4e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa717d1e209131cea51970.md>)

[Implementation](<https://llava-vl.github.io/>)

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

### LLaVA

Grounding: 41.7. Result ID: 6991b4c33a6c3148490a0d75.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71201e209131cea4f4e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa717d1e209131cea51970.md>)

[Implementation](<https://llava-vl.github.io/>)

Recorded evaluation configuration: {"tool\_configuration":"Tools in Test"}. This record does not assert independent reproduction.

### Bard

Grounding: 36.5. Result ID: 6991b4c33a6c3148490a0d78.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71201e209131cea4f4e0.md>)

Recorded evaluation configuration: {"evaluation\_date":"0730"}. This record does not assert independent reproduction.

### MM-REACT

Grounding: 30.2. Result ID: 6991b4c33a6c3148490a0d7a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71201e209131cea4f4e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ed31e209131cea4272b.md>)

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

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