# Multimodal Hallucination Evaluation on HallBench

Source: [Multimodal Hallucination Evaluation on HallBench](<https://www.sota2.com/research/sota/multimodal-hallucination-evaluation-on-hallbench>)

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

Data updated: 2026-03-31T09:20:43.372Z

Evaluation of model performance on multimodal hallucination detection using the HallBench dataset, typically measured by Score.

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 Hallucination Evaluation |
| Dataset (including any stated split) | HallBench |
| Metric | Score |
| Unit | SCALAR |
| 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

- [Score](<https://www.sota2.com/research/sota/multimodal-hallucination-evaluation-on-hallbench.md?metric=Score>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Score | Result date | Report paper |
| --- | --- | --- | --- |
| GPT-5-Thinking | 65.2 | Not recorded | 68fa7bd11e209131cea8ab14 |
| Gemini-2.5-Pro | 64.1 | Not recorded | 68fa7bd11e209131cea8ab14 |
| VAPO-Thinker-7B | 57.4 | Not recorded | 68fa7bd11e209131cea8ab14 |
| VLAA-Thinker-7B | 54.7 | Not recorded | 68fa7bd11e209131cea8ab14 |
| R1-OneVision-7B | 52.5 | Not recorded | 68fa7bd11e209131cea8ab14 |
| Qwen2.5-VL-7B | 50 | Not recorded | 68fa7bd11e209131cea8ab14 |
| Vision-R1-7B | 49.5 | Not recorded | 68fa7bd11e209131cea8ab14 |
| VAPO-Thinker-3B | 49.5 | Not recorded | 68fa7bd11e209131cea8ab14 |
| InternVL2.5-8B | 49 | Not recorded | 68fa7bd11e209131cea8ab14 |

### GPT-5-Thinking

Score: 65.2. Result ID: 69cb909547f34ce924b41b84.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bd11e209131cea8ab14.md>)

Recorded evaluation configuration: {"Model Category":"Close-source Models"}. This record does not assert independent reproduction.

### Gemini-2.5-Pro

Score: 64.1. Result ID: 69cb909547f34ce924b41b85.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bd11e209131cea8ab14.md>)

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

Recorded evaluation configuration: {"Model Category":"Close-source Models"}. This record does not assert independent reproduction.

### VAPO-Thinker-7B

Score: 57.4. Result ID: 69cb909547f34ce924b41b8c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bd11e209131cea8ab14.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bd11e209131cea8ab14.md>)

Recorded evaluation configuration: {"Model Category":"Our Models"}. This record does not assert independent reproduction.

### VLAA-Thinker-7B

Score: 54.7. Result ID: 69cb909547f34ce924b41b89.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bd11e209131cea8ab14.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78051e209131cea765df.md>)

Recorded evaluation configuration: {"Model Category":"Open-source Models"}. This record does not assert independent reproduction.

### R1-OneVision-7B

Score: 52.5. Result ID: 69cb909547f34ce924b41b88.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bd11e209131cea8ab14.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77771e209131cea7357a.md>)

Recorded evaluation configuration: {"Model Category":"Open-source Models"}. This record does not assert independent reproduction.

### Qwen2.5-VL-7B

Score: 50. Result ID: 69cb909547f34ce924b41b86.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bd11e209131cea8ab14.md>)

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

Recorded evaluation configuration: {"Model Category":"Open-source Models"}. This record does not assert independent reproduction.

### Vision-R1-7B

Score: 49.5. Result ID: 69cb909547f34ce924b41b8a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bd11e209131cea8ab14.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77591e209131cea72979.md>)

Recorded evaluation configuration: {"Model Category":"Open-source Models"}. This record does not assert independent reproduction.

### VAPO-Thinker-3B

Score: 49.5. Result ID: 69cb909547f34ce924b41b8b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bd11e209131cea8ab14.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bd11e209131cea8ab14.md>)

Recorded evaluation configuration: {"Model Category":"Our Models"}. This record does not assert independent reproduction.

### InternVL2.5-8B

Score: 49. Result ID: 69cb909547f34ce924b41b87.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bd11e209131cea8ab14.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bab1e209131cea89c7f.md>)

[Implementation](<https://github.com/OpenGVLab/InternVL>)

Recorded evaluation configuration: {"Model Category":"Open-source Models"}. This record does not assert independent reproduction.

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