# Text-to-Video Generation on CountBench Source: [Text-to-Video Generation on CountBench]() Collection index: [llms.txt]() Data updated: 2026-04-10T09:56:30.236Z Evaluation of Text-to-Video Generation models using the CountBench dataset, assessing metrics related to object counting accuracy, temporal consistency (TC), and visual fidelity (CLIP Score). 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 | Text-to-Video Generation | | Dataset (including any stated split) | CountBench | | Metric | Counting Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 16 of 16 | | 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 - [Counting Accuracy](): Higher is better; unit: PERCENT. - [Temporal Consistency (TC)](): Higher is better; unit: PERCENT. - [CLIP Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Counting Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Wan2.1-14B + NUMINA | 0.591 | Not recorded | 69d8841d7d0172140ec77a79 | | Wan2.1-14B + Prompt enhancement | 0.569 | Not recorded | 69d8841d7d0172140ec77a79 | | Wan2.1-14B + Seed search | 0.561 | Not recorded | 69d8841d7d0172140ec77a79 | | Wan2.1-14B | 0.536 | Not recorded | 69d8841d7d0172140ec77a79 | | Wan2.2-5B + NUMINA | 0.527 | Not recorded | 69d8841d7d0172140ec77a79 | | Wan2.1-1.3B + NUMINA | 0.497 | Not recorded | 69d8841d7d0172140ec77a79 | | Wan2.2-5B + Prompt enhancement | 0.49 | Not recorded | 69d8841d7d0172140ec77a79 | | Wan2.2-5B + Seed search | 0.488 | Not recorded | 69d8841d7d0172140ec77a79 | | Wan2.2-5B | 0.478 | Not recorded | 69d8841d7d0172140ec77a79 | | Wan2.1-1.3B + Prompt enhancement | 0.472 | Not recorded | 69d8841d7d0172140ec77a79 | | Wan2.1-1.3B + Seed search | 0.455 | Not recorded | 69d8841d7d0172140ec77a79 | | NUMINA | 0.444 | Not recorded | 69d8841d7d0172140ec77a79 | | CogVideoX-5B + Seed search | 0.427 | Not recorded | 69d8841d7d0172140ec77a79 | | CogVideoX-5B + Prompt enhancement | 0.425 | Not recorded | 69d8841d7d0172140ec77a79 | | Wan2.1-1.3B | 0.423 | Not recorded | 69d8841d7d0172140ec77a79 | | CogVideoX-5B | 0.402 | Not recorded | 69d8841d7d0172140ec77a79 | ### Wan2.1-14B + NUMINA Counting Accuracy: 0.591. Result ID: 69d8c94147f34ce924c22ba9. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model scale":"14B","Frames":"81","Resolution":"1280×720","Strategy":"NUMINA"}. This record does not assert independent reproduction. ### Wan2.1-14B + Prompt enhancement Counting Accuracy: 0.569. Result ID: 69d8c94147f34ce924c22ba8. [Reporting paper]() Recorded evaluation configuration: {"Model scale":"14B","Frames":"81","Resolution":"1280×720","Strategy":"Prompt enhancement"}. This record does not assert independent reproduction. ### Wan2.1-14B + Seed search Counting Accuracy: 0.561. Result ID: 69d8c94147f34ce924c22ba7. [Reporting paper]() Recorded evaluation configuration: {"Model scale":"14B","Frames":"81","Resolution":"1280×720","Strategy":"Seed search"}. This record does not assert independent reproduction. ### Wan2.1-14B Counting Accuracy: 0.536. Result ID: 69d8c94147f34ce924c22ba6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model scale":"14B","Frames":"81","Resolution":"1280×720","Strategy":"Base"}. This record does not assert independent reproduction. ### Wan2.2-5B + NUMINA Counting Accuracy: 0.527. Result ID: 69d8c94147f34ce924c22ba5. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model scale":"5B","Frames":"81","Resolution":"1280×704","Strategy":"NUMINA"}. This record does not assert independent reproduction. ### Wan2.1-1.3B + NUMINA Counting Accuracy: 0.497. Result ID: 69d8c94147f34ce924c22ba1. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model scale":"1.3B","Frames":"81","Resolution":"832×480","Strategy":"NUMINA"}. This record does not assert independent reproduction. ### Wan2.2-5B + Prompt enhancement Counting Accuracy: 0.49. Result ID: 69d8c94147f34ce924c22ba4. [Reporting paper]() Recorded evaluation configuration: {"Model scale":"5B","Frames":"81","Resolution":"1280×704","Strategy":"Prompt enhancement"}. This record does not assert independent reproduction. ### Wan2.2-5B + Seed search Counting Accuracy: 0.488. Result ID: 69d8c94147f34ce924c22ba3. [Reporting paper]() Recorded evaluation configuration: {"Model scale":"5B","Frames":"81","Resolution":"1280×704","Strategy":"Seed search"}. This record does not assert independent reproduction. ### Wan2.2-5B Counting Accuracy: 0.478. Result ID: 69d8c94147f34ce924c22ba2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model scale":"5B","Frames":"81","Resolution":"1280×704","Strategy":"Base"}. This record does not assert independent reproduction. ### Wan2.1-1.3B + Prompt enhancement Counting Accuracy: 0.472. Result ID: 69d8c94147f34ce924c22ba0. [Reporting paper]() Recorded evaluation configuration: {"Model scale":"1.3B","Frames":"81","Resolution":"832×480","Strategy":"Prompt enhancement"}. This record does not assert independent reproduction. ### Wan2.1-1.3B + Seed search Counting Accuracy: 0.455. Result ID: 69d8c94147f34ce924c22b9f. [Reporting paper]() Recorded evaluation configuration: {"Model scale":"1.3B","Frames":"81","Resolution":"832×480","Strategy":"Seed search"}. This record does not assert independent reproduction. ### NUMINA Counting Accuracy: 0.444. Result ID: 69d8c94147f34ce924c22b9d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"CogVideoX-5B","Frames":"81","Resolution":"1360×768"}. This record does not assert independent reproduction. ### CogVideoX-5B + Seed search Counting Accuracy: 0.427. Result ID: 69d8c94147f34ce924c22b9b. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"CogVideoX-5B","Frames":"81","Resolution":"1360×768","Seed search":"true"}. This record does not assert independent reproduction. ### CogVideoX-5B + Prompt enhancement Counting Accuracy: 0.425. Result ID: 69d8c94147f34ce924c22b9c. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"CogVideoX-5B","Frames":"81","Resolution":"1360×768","Prompt enhancement":"true"}. This record does not assert independent reproduction. ### Wan2.1-1.3B Counting Accuracy: 0.423. Result ID: 69d8c94147f34ce924c22b9e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model scale":"1.3B","Frames":"81","Resolution":"832×480","Strategy":"Base"}. This record does not assert independent reproduction. ### CogVideoX-5B Counting Accuracy: 0.402. Result ID: 69d8c94147f34ce924c22b9a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"CogVideoX-5B","Frames":"81","Resolution":"1360×768"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.