# General Image Reasoning and Understanding on General Reasoning and Perception Aggregates

Source: [General Image Reasoning and Understanding on General Reasoning and Perception Aggregates](<https://www.sota2.com/research/sota/general-image-reasoning-and-understanding-on-general-reasoning-and-perception-aggregates>)

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

Data updated: 2026-03-20T08:33:13.039Z

Evaluation of model performance on general image reasoning and understanding capabilities using the General Reasoning and Perception Aggregates dataset, reporting separate scores for Reasoning and Perception components.

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 | General Image Reasoning and Understanding |
| Dataset (including any stated split) | General Reasoning and Perception Aggregates |
| Metric | Average Reasoning Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 12 of 12 |
| 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

- [Average Reasoning Score](<https://www.sota2.com/research/sota/general-image-reasoning-and-understanding-on-general-reasoning-and-perception-aggregates.md?metric=Average%20Reasoning%20Score>): Higher is better; unit: PERCENT.

- [Average Perception Score](<https://www.sota2.com/research/sota/general-image-reasoning-and-understanding-on-general-reasoning-and-perception-aggregates.md?metric=Average%20Perception%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Average Reasoning Score | Result date | Report paper |
| --- | --- | --- | --- |
| Insight-V++ | 71.1 | Not recorded | 69bcd45d0636e5bf4afc27b3 |
| Qwen2.5-VL + Multi-Agent (RL) | 69.5 | Not recorded | 69bcd45d0636e5bf4afc27b3 |
| Qwen2.5-VL | 66.3 | Not recorded | 69bcd45d0636e5bf4afc27b3 |
| Insight-V | 60.1 | Not recorded | 69bcd45d0636e5bf4afc27b3 |
| Our Base Model + Multi-Agent | 59 | Not recorded | 69bcd45d0636e5bf4afc27b3 |
| Ovis1.5-LLaMA3 | 57.5 | Not recorded | 69bcd45d0636e5bf4afc27b3 |
| Our Base Model | 56.8 | Not recorded | 69bcd45d0636e5bf4afc27b3 |
| Idefics3-LLaMA3 | 56 | Not recorded | 69bcd45d0636e5bf4afc27b3 |
| Insight-V-LLaVA | 54.9 | Not recorded | 69bcd45d0636e5bf4afc27b3 |
| LLaVA-NeXT-LLaMA3 + Multi-Agent | 51.8 | Not recorded | 69bcd45d0636e5bf4afc27b3 |
| LLaVA-NeXT-LLaMA3 | 46.8 | Not recorded | 69bcd45d0636e5bf4afc27b3 |

### Insight-V++

Average Reasoning Score: 71.1. Result ID: 69bd048947f34ce924a704ea.

[Reporting paper](<https://www.sota2.com/research/paper/69bcd45d0636e5bf4afc27b3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69bcd45d0636e5bf4afc27b3.md>)

Recorded evaluation configuration: {"Size":"7B","Base Model":"Qwen2.5-VL","Self-Evolving":"true"}. This record does not assert independent reproduction.

### Qwen2.5-VL + Multi-Agent (RL)

Average Reasoning Score: 69.5. Result ID: 69bd048947f34ce924a704e9.

[Reporting paper](<https://www.sota2.com/research/paper/69bcd45d0636e5bf4afc27b3.md>)

Recorded evaluation configuration: {"Size":"7B","Multi-Agent (RL)":"true"}. This record does not assert independent reproduction.

### Qwen2.5-VL

Average Reasoning Score: 66.3. Result ID: 69bd048947f34ce924a704e8.

[Reporting paper](<https://www.sota2.com/research/paper/69bcd45d0636e5bf4afc27b3.md>)

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

Recorded evaluation configuration: {"Size":"7B"}. This record does not assert independent reproduction.

### Insight-V

Average Reasoning Score: 60.1. Result ID: 69bd048947f34ce924a704e7.

[Reporting paper](<https://www.sota2.com/research/paper/69bcd45d0636e5bf4afc27b3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69bcd45d0636e5bf4afc27b3.md>)

Recorded evaluation configuration: {"Size":"7B","Base Model":"Our Base Model","Iterative DPO":"true"}. This record does not assert independent reproduction.

### Our Base Model + Multi-Agent

Average Reasoning Score: 59. Result ID: 69bd048947f34ce924a704e6.

[Reporting paper](<https://www.sota2.com/research/paper/69bcd45d0636e5bf4afc27b3.md>)

Recorded evaluation configuration: {"Size":"7B","Multi-Agent":"true"}. This record does not assert independent reproduction.

### Ovis1.5-LLaMA3

Average Reasoning Score: 57.5. Result ID: 69bd048947f34ce924a704e1.

[Reporting paper](<https://www.sota2.com/research/paper/69bcd45d0636e5bf4afc27b3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73b21e209131cea5dfaa.md>)

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

### Our Base Model

Average Reasoning Score: 56.8. Result ID: 69bd048947f34ce924a704e5.

[Reporting paper](<https://www.sota2.com/research/paper/69bcd45d0636e5bf4afc27b3.md>)

Recorded evaluation configuration: {"Size":"7B"}. This record does not assert independent reproduction.

### Idefics3-LLaMA3

Average Reasoning Score: 56. Result ID: 69bd048947f34ce924a704e0.

[Reporting paper](<https://www.sota2.com/research/paper/69bcd45d0636e5bf4afc27b3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74931e209131cea629e0.md>)

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

### Insight-V-LLaVA

Average Reasoning Score: 54.9. Result ID: 69bd048947f34ce924a704e4.

[Reporting paper](<https://www.sota2.com/research/paper/69bcd45d0636e5bf4afc27b3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69bcd45d0636e5bf4afc27b3.md>)

Recorded evaluation configuration: {"Size":"8B","Base Model":"LLaVA-NeXT-LLaMA3","Iterative DPO":"true"}. This record does not assert independent reproduction.

### LLaVA-NeXT-LLaMA3 + Multi-Agent

Average Reasoning Score: 51.8. Result ID: 69bd048947f34ce924a704e3.

[Reporting paper](<https://www.sota2.com/research/paper/69bcd45d0636e5bf4afc27b3.md>)

Recorded evaluation configuration: {"Size":"8B","Multi-Agent":"true"}. This record does not assert independent reproduction.

### LLaVA-NeXT-LLaMA3

Average Reasoning Score: 46.8. Result ID: 69bd048947f34ce924a704e2.

[Reporting paper](<https://www.sota2.com/research/paper/69bcd45d0636e5bf4afc27b3.md>)

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

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