# Image Reasoning on ZeroBench sub

Source: [Image Reasoning on ZeroBench sub](<https://www.sota2.com/research/sota/image-reasoning-on-zerobench-sub>)

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

Data updated: 2026-05-01T07:23:01.673Z

Evaluation of model performance on image reasoning tasks using the ZeroBench sub dataset, measured using Accuracy.

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 | Image Reasoning |
| Dataset (including any stated split) | ZeroBench sub |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 14 of 14 |
| 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

- [Accuracy](<https://www.sota2.com/research/sota/image-reasoning-on-zerobench-sub.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| CoPD | 24.4 | Not recorded | 69f4338b55a3c32af89fd0fc |
| CoPD | 23.35 | Not recorded | 69f4338b55a3c32af89fd0fc |
| MOPD | 22.75 | Not recorded | 69f4338b55a3c32af89fd0fc |
| Mixed RLVR | 22.6 | Not recorded | 69f4338b55a3c32af89fd0fc |
| Text-Expert | 22.01 | Not recorded | 69f4338b55a3c32af89fd0fc |
| Text-Expert | 22.01 | Not recorded | 69f4338b55a3c32af89fd0fc |
| Video-Expert | 22.01 | Not recorded | 69f4338b55a3c32af89fd0fc |
| Mixed RLVR | 21.86 | Not recorded | 69f4338b55a3c32af89fd0fc |
| OPD T→V | 21.71 | Not recorded | 69f4338b55a3c32af89fd0fc |
| Base | 21.41 | Not recorded | 69f4338b55a3c32af89fd0fc |
| Image-Expert | 21.41 | Not recorded | 69f4338b55a3c32af89fd0fc |
| OPD V→T | 21.41 | Not recorded | 69f4338b55a3c32af89fd0fc |
| Base | 21.41 | Not recorded | 69f4338b55a3c32af89fd0fc |
| Image-Expert | 21.41 | Not recorded | 69f4338b55a3c32af89fd0fc |

### CoPD

Accuracy: 24.4. Result ID: 69f4544a47f34ce924d9d1b5.

[Reporting paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

Recorded evaluation configuration: {"Training Strategy":"Co-Evolving Policy Distillation"}. This record does not assert independent reproduction.

### CoPD

Accuracy: 23.35. Result ID: 69f4544a47f34ce924d9d209.

[Reporting paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

Recorded evaluation configuration: {"Training Strategy":"CoPD"}. This record does not assert independent reproduction.

### MOPD

Accuracy: 22.75. Result ID: 69f4544a47f34ce924d9d208.

[Reporting paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

[Method origin paper](<https://www.sota2.com/research/paper/695de86653fdd5c3617439f3.md>)

[Implementation](<https://github.com/XiaomiMiMo/MiMo-V2-Flash>)

Recorded evaluation configuration: {"Training Strategy":"MOPD"}. This record does not assert independent reproduction.

### Mixed RLVR

Accuracy: 22.6. Result ID: 69f4544a47f34ce924d9d207.

[Reporting paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

Recorded evaluation configuration: {"Training Strategy":"Mixed RLVR"}. This record does not assert independent reproduction.

### Text-Expert

Accuracy: 22.01. Result ID: 69f4544a47f34ce924d9d1b1.

[Reporting paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

Recorded evaluation configuration: {"Training Strategy":"Text-Specific Expert"}. This record does not assert independent reproduction.

### Text-Expert

Accuracy: 22.01. Result ID: 69f4544a47f34ce924d9d205.

[Reporting paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

Recorded evaluation configuration: {"Training Strategy":"Text-Expert"}. This record does not assert independent reproduction.

### Video-Expert

Accuracy: 22.01. Result ID: 69f4544a47f34ce924d9d206.

[Reporting paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

Recorded evaluation configuration: {"Training Strategy":"Video-Expert"}. This record does not assert independent reproduction.

### Mixed RLVR

Accuracy: 21.86. Result ID: 69f4544a47f34ce924d9d1b2.

[Reporting paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

Recorded evaluation configuration: {"Training Strategy":"Joint optimization on combined budget"}. This record does not assert independent reproduction.

### OPD T→V

Accuracy: 21.71. Result ID: 69f4544a47f34ce924d9d1b4.

[Reporting paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

Recorded evaluation configuration: {"Training Strategy":"Static On-policy Policy Distillation (Text to Image)"}. This record does not assert independent reproduction.

### Base

Accuracy: 21.41. Result ID: 69f4544a47f34ce924d9d1af.

[Reporting paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

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

Recorded evaluation configuration: {"Training Strategy":"Base Model"}. This record does not assert independent reproduction.

### Image-Expert

Accuracy: 21.41. Result ID: 69f4544a47f34ce924d9d1b0.

[Reporting paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

Recorded evaluation configuration: {"Training Strategy":"Image-Specific Expert"}. This record does not assert independent reproduction.

### OPD V→T

Accuracy: 21.41. Result ID: 69f4544a47f34ce924d9d1b3.

[Reporting paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

Recorded evaluation configuration: {"Training Strategy":"Static On-policy Policy Distillation (Image to Text)"}. This record does not assert independent reproduction.

### Base

Accuracy: 21.41. Result ID: 69f4544a47f34ce924d9d203.

[Reporting paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

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

Recorded evaluation configuration: {"Training Strategy":"Base"}. This record does not assert independent reproduction.

### Image-Expert

Accuracy: 21.41. Result ID: 69f4544a47f34ce924d9d204.

[Reporting paper](<https://www.sota2.com/research/paper/69f4338b55a3c32af89fd0fc.md>)

Recorded evaluation configuration: {"Training Strategy":"Image-Expert"}. This record does not assert independent reproduction.

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