# Video Generation on EVD-Bench

Source: [Video Generation on EVD-Bench](<https://www.sota2.com/research/sota/video-generation-on-evd-bench>)

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

Data updated: 2026-03-17T08:00:10.055Z

Evaluation of video generation models on the EVD-Bench dataset, assessed using human evaluations for text fidelity, quality, and dynamics, alongside automated metrics for appearance and dynamics.

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 | Video Generation |
| Dataset (including any stated split) | EVD-Bench |
| Metric | Human Eval: Text Fidelity |
| Unit | SCALAR |
| 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

- [Human Eval: Text Fidelity](<https://www.sota2.com/research/sota/video-generation-on-evd-bench.md?metric=Human%20Eval%3A%20Text%20Fidelity>): Higher is better; unit: SCALAR.

- [Human Eval: Quality](<https://www.sota2.com/research/sota/video-generation-on-evd-bench.md?metric=Human%20Eval%3A%20Quality>): Higher is better; unit: SCALAR.

- [Human Eval: Dynamics](<https://www.sota2.com/research/sota/video-generation-on-evd-bench.md?metric=Human%20Eval%3A%20Dynamics>): Higher is better; unit: SCALAR.

- [Automated Metrics: Appearance](<https://www.sota2.com/research/sota/video-generation-on-evd-bench.md?metric=Automated%20Metrics%3A%20Appearance>): Higher is better; unit: SCALAR.

- [Automated Metrics: Dynamics](<https://www.sota2.com/research/sota/video-generation-on-evd-bench.md?metric=Automated%20Metrics%3A%20Dynamics>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Human Eval: Text Fidelity | Result date | Report paper |
| --- | --- | --- | --- |
| DiT-30B + EVD | 89.7 | Not recorded | 69b8e00a0636e5bf4afbea06 |
| DiT-4B + EVD | 88.9 | Not recorded | 69b8e00a0636e5bf4afbea06 |
| CogVideo2B | 80.2 | Not recorded | 69b8e00a0636e5bf4afbea06 |
| PyramidFlow | 75.8 | Not recorded | 69b8e00a0636e5bf4afbea06 |
| DiT-30B | 72.4 | Not recorded | 69b8e00a0636e5bf4afbea06 |
| DiT-4B | 70.6 | Not recorded | 69b8e00a0636e5bf4afbea06 |
| Veo 3.1 | 66.9 | Not recorded | 69b8e00a0636e5bf4afbea06 |
| CogVideo5B | 65.4 | Not recorded | 69b8e00a0636e5bf4afbea06 |
| Runway Gen-4.5 | 64.7 | Not recorded | 69b8e00a0636e5bf4afbea06 |
| Sora 2 Pro | 63.5 | Not recorded | 69b8e00a0636e5bf4afbea06 |
| Kling 3.0 Pro | 61.8 | Not recorded | 69b8e00a0636e5bf4afbea06 |
| Mochi 1 | 58.2 | Not recorded | 69b8e00a0636e5bf4afbea06 |

### DiT-30B + EVD

Human Eval: Text Fidelity: 89.7. Result ID: 69b909b647f34ce924a15ab4.

[Reporting paper](<https://www.sota2.com/research/paper/69b8e00a0636e5bf4afbea06.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b8e00a0636e5bf4afbea06.md>)

Recorded evaluation configuration: {"number of parameters":"30B","module":"EVD"}. This record does not assert independent reproduction.

### DiT-4B + EVD

Human Eval: Text Fidelity: 88.9. Result ID: 69b909b647f34ce924a15ab9.

[Reporting paper](<https://www.sota2.com/research/paper/69b8e00a0636e5bf4afbea06.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b8e00a0636e5bf4afbea06.md>)

Recorded evaluation configuration: {"Parameters":"4B","EVD integration":"true"}. This record does not assert independent reproduction.

### CogVideo2B

Human Eval: Text Fidelity: 80.2. Result ID: 69b909b647f34ce924a15ab5.

[Reporting paper](<https://www.sota2.com/research/paper/69b8e00a0636e5bf4afbea06.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77a01e209131cea74741.md>)

Recorded evaluation configuration: {"Parameters":"2B"}. This record does not assert independent reproduction.

### PyramidFlow

Human Eval: Text Fidelity: 75.8. Result ID: 69b909b647f34ce924a15ab7.

[Reporting paper](<https://www.sota2.com/research/paper/69b8e00a0636e5bf4afbea06.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa777b1e209131cea737d3.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### DiT-30B

Human Eval: Text Fidelity: 72.4. Result ID: 69b909b647f34ce924a15ab3.

[Reporting paper](<https://www.sota2.com/research/paper/69b8e00a0636e5bf4afbea06.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ea91e209131cea41328.md>)

Recorded evaluation configuration: {"number of parameters":"30B"}. This record does not assert independent reproduction.

### DiT-4B

Human Eval: Text Fidelity: 70.6. Result ID: 69b909b647f34ce924a15ab8.

[Reporting paper](<https://www.sota2.com/research/paper/69b8e00a0636e5bf4afbea06.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ea91e209131cea41328.md>)

Recorded evaluation configuration: {"Parameters":"4B"}. This record does not assert independent reproduction.

### Veo 3.1

Human Eval: Text Fidelity: 66.9. Result ID: 69b909b647f34ce924a15ab0.

[Reporting paper](<https://www.sota2.com/research/paper/69b8e00a0636e5bf4afbea06.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### CogVideo5B

Human Eval: Text Fidelity: 65.4. Result ID: 69b909b647f34ce924a15ab6.

[Reporting paper](<https://www.sota2.com/research/paper/69b8e00a0636e5bf4afbea06.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77a01e209131cea74741.md>)

Recorded evaluation configuration: {"Parameters":"5B"}. This record does not assert independent reproduction.

### Runway Gen-4.5

Human Eval: Text Fidelity: 64.7. Result ID: 69b909b647f34ce924a15aaf.

[Reporting paper](<https://www.sota2.com/research/paper/69b8e00a0636e5bf4afbea06.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Sora 2 Pro

Human Eval: Text Fidelity: 63.5. Result ID: 69b909b647f34ce924a15ab1.

[Reporting paper](<https://www.sota2.com/research/paper/69b8e00a0636e5bf4afbea06.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Kling 3.0 Pro

Human Eval: Text Fidelity: 61.8. Result ID: 69b909b647f34ce924a15aae.

[Reporting paper](<https://www.sota2.com/research/paper/69b8e00a0636e5bf4afbea06.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Mochi 1

Human Eval: Text Fidelity: 58.2. Result ID: 69b909b647f34ce924a15ab2.

[Reporting paper](<https://www.sota2.com/research/paper/69b8e00a0636e5bf4afbea06.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

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