# Vision-Language Modeling on LiveB, LWilder, and LCOCO Combined

Source: [Vision-Language Modeling on LiveB, LWilder, and LCOCO Combined](<https://www.sota2.com/research/sota/vision-language-modeling-on-liveb-lwilder-and-lcoco-combined>)

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

Data updated: 2026-02-26T05:36:20.591Z

Evaluation of Vision-Language Modeling performance across a combined dataset consisting of LiveB, LWilder, and LCOCO.

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 | Vision-Language Modeling |
| Dataset (including any stated split) | LiveB, LWilder, and LCOCO Combined |
| Metric | Avg Relative Improvement |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [Avg Relative Improvement](<https://www.sota2.com/research/sota/vision-language-modeling-on-liveb-lwilder-and-lcoco-combined.md?metric=Avg%20Relative%20Improvement>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Avg Relative Improvement | Result date | Report paper |
| --- | --- | --- | --- |
| CASP\_GPTQ | 20.1 | Not recorded | 68fa775f1e209131cea72cc3 |
| CASP\_AQLM | 3.4 | Not recorded | 68fa775f1e209131cea72cc3 |
| CASP\_QuIP# | 0.01 | Not recorded | 68fa775f1e209131cea72cc3 |

### CASP\_GPTQ

Avg Relative Improvement: 20.1. Result ID: 699499fc3a6c3148491e5260.

[Reporting paper](<https://www.sota2.com/research/paper/68fa775f1e209131cea72cc3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa775f1e209131cea72cc3.md>)

[Implementation](<https://github.com/vbdi/casp>)

Recorded evaluation configuration: {"Bit":"3","Base Model":"LLaVA1.5-7B"}. This record does not assert independent reproduction.

### CASP\_AQLM

Avg Relative Improvement: 3.4. Result ID: 699499fc3a6c3148491e5261.

[Reporting paper](<https://www.sota2.com/research/paper/68fa775f1e209131cea72cc3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa775f1e209131cea72cc3.md>)

[Implementation](<https://github.com/vbdi/casp>)

Recorded evaluation configuration: {"Bit":"3","Base Model":"LLaVA1.5-7B"}. This record does not assert independent reproduction.

### CASP\_QuIP#

Avg Relative Improvement: 0.01. Result ID: 699499fc3a6c3148491e5262.

[Reporting paper](<https://www.sota2.com/research/paper/68fa775f1e209131cea72cc3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa775f1e209131cea72cc3.md>)

[Implementation](<https://github.com/vbdi/casp>)

Recorded evaluation configuration: {"Bit":"3","Base Model":"LLaVA1.5-7B"}. This record does not assert independent reproduction.

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