# Multimodal Action Labeling on CrossModal-CS 50 paired tasks

Source: [Multimodal Action Labeling on CrossModal-CS 50 paired tasks](<https://www.sota2.com/research/sota/multimodal-action-labeling-on-crossmodal-cs-50-paired-tasks>)

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

Data updated: 2026-04-15T08:04:09.173Z

Evaluation of multimodal action labeling performance using the CrossModal-CS 50 paired tasks dataset, reporting metrics such as TCA (%), Latency (s), BW (KB/task), and Native routing (%).

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 Action Labeling |
| Dataset (including any stated split) | CrossModal-CS 50 paired tasks |
| Metric | TCA |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 2 of 2 |
| 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

- [TCA](<https://www.sota2.com/research/sota/multimodal-action-labeling-on-crossmodal-cs-50-paired-tasks.md?metric=TCA>): Higher is better; unit: PERCENT.

- [Latency (s)](<https://www.sota2.com/research/sota/multimodal-action-labeling-on-crossmodal-cs-50-paired-tasks.md?metric=Latency%20(s)>): Lower is better; unit: TIME.

- [BW (KB/task)](<https://www.sota2.com/research/sota/multimodal-action-labeling-on-crossmodal-cs-50-paired-tasks.md?metric=BW%20(KB%2Ftask)>): Lower is better; unit: SIZE.

- [Native routing](<https://www.sota2.com/research/sota/multimodal-action-labeling-on-crossmodal-cs-50-paired-tasks.md?metric=Native%20routing>): Higher is better; unit: PERCENT.

## Recorded results

| Method | TCA | Result date | Report paper |
| --- | --- | --- | --- |
| MMA2A | 52 | Not recorded | 69df1ba67d0172140ec7ba21 |
| Text-BN | 32 | Not recorded | 69df1ba67d0172140ec7ba21 |

### MMA2A

TCA: 52. Result ID: 69df466a47f34ce924c7df74.

[Reporting paper](<https://www.sota2.com/research/paper/69df1ba67d0172140ec7ba21.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69df1ba67d0172140ec7ba21.md>)

Recorded evaluation configuration: {"Routing Strategy":"Modality-native routing","Model":"Gemini 2.5 Flash"}. This record does not assert independent reproduction.

### Text-BN

TCA: 32. Result ID: 69df466a47f34ce924c7df73.

[Reporting paper](<https://www.sota2.com/research/paper/69df1ba67d0172140ec7ba21.md>)

Recorded evaluation configuration: {"Routing Strategy":"Text-Bottleneck","Model":"Gemini 2.5 Flash"}. This record does not assert independent reproduction.

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