# Report Generation on MIMIC-CXR-JPG (test) Source: [Report Generation on MIMIC-CXR-JPG (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:37:18.512Z Evaluation of models generating radiology reports from the test split of the MIMIC-CXR-JPG dataset, measured using standard text generation metrics like BLEU and ROUGE. 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 | Report Generation | | Dataset (including any stated split) | MIMIC-CXR-JPG (test) | | Metric | BLEU-4 | | Unit | SCALAR | | 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 - [BLEU-4](): Higher is better; unit: SCALAR. - [ROUGE](): Higher is better; unit: SCALAR. - [GREEN](): Higher is better; unit: OTHER. ## Recorded results | Method | BLEU-4 | Result date | Report paper | | --- | --- | --- | --- | | VILA-M3-13B | 21.6 | Not recorded | 68fa773a1e209131cea71cf8 | | VILA-M3-40B | 21.6 | Not recorded | 68fa773a1e209131cea71cf8 | | VILA-M3-40B | 21.6 | Not recorded | 68fa773a1e209131cea71cf8 | | VILA-M3-8B | 21.4 | Not recorded | 68fa773a1e209131cea71cf8 | | VILA-M3-13B | 21.4 | Not recorded | 68fa773a1e209131cea71cf8 | | VILA-M3-8B | 21.2 | Not recorded | 68fa773a1e209131cea71cf8 | | Med-Gemini | 20.5 | Not recorded | 68fa773a1e209131cea71cf8 | | Med-Gemini | 20.5 | Not recorded | 68fa773a1e209131cea71cf8 | | VILA-M3-3B | 20.2 | Not recorded | 68fa773a1e209131cea71cf8 | | VILA-M3-3B | 19.7 | Not recorded | 68fa773a1e209131cea71cf8 | | Llava-Rad | 15.4 | Not recorded | 68fa773a1e209131cea71cf8 | | Llava-Rad | 15.4 | Not recorded | 68fa773a1e209131cea71cf8 | | DCL | 10.9 | Not recorded | 68fa773a1e209131cea71cf8 | | DCL | 10.9 | Not recorded | 68fa773a1e209131cea71cf8 | | Llava-Med | 1 | Not recorded | 68fa773a1e209131cea71cf8 | | Llava-Med | 1 | Not recorded | 68fa773a1e209131cea71cf8 | ### VILA-M3-13B BLEU-4: 21.6. Result ID: 69946e8d3a6c3148491d33d4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"13B","expert\_knowledge\_inclusion":true}. This record does not assert independent reproduction. ### VILA-M3-40B BLEU-4: 21.6. Result ID: 69946e8d3a6c3148491d33d5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"40B","expert\_knowledge\_inclusion":false}. This record does not assert independent reproduction. ### VILA-M3-40B BLEU-4: 21.6. Result ID: 69946e8d3a6c3148491d33d6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"40B","expert\_knowledge\_inclusion":true}. This record does not assert independent reproduction. ### VILA-M3-8B BLEU-4: 21.4. Result ID: 69946e8d3a6c3148491d33d1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"8B","expert\_knowledge\_inclusion":false}. This record does not assert independent reproduction. ### VILA-M3-13B BLEU-4: 21.4. Result ID: 69946e8d3a6c3148491d33d3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"13B","expert\_knowledge\_inclusion":false}. This record does not assert independent reproduction. ### VILA-M3-8B BLEU-4: 21.2. Result ID: 69946e8d3a6c3148491d33d2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"8B","expert\_knowledge\_inclusion":true}. This record does not assert independent reproduction. ### Med-Gemini BLEU-4: 20.5. Result ID: 69946e8d3a6c3148491d33c7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"1.5T","expert\_knowledge\_inclusion":false}. This record does not assert independent reproduction. ### Med-Gemini BLEU-4: 20.5. Result ID: 69946e8d3a6c3148491d33c8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"1.5T","expert\_knowledge\_inclusion":true}. This record does not assert independent reproduction. ### VILA-M3-3B BLEU-4: 20.2. Result ID: 69946e8d3a6c3148491d33d0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"3B","expert\_knowledge\_inclusion":true}. This record does not assert independent reproduction. ### VILA-M3-3B BLEU-4: 19.7. Result ID: 69946e8d3a6c3148491d33cf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"3B","expert\_knowledge\_inclusion":false}. This record does not assert independent reproduction. ### Llava-Rad BLEU-4: 15.4. Result ID: 69946e8d3a6c3148491d33cb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"7B","expert\_knowledge\_inclusion":false}. This record does not assert independent reproduction. ### Llava-Rad BLEU-4: 15.4. Result ID: 69946e8d3a6c3148491d33cc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Params":"7B","expert\_knowledge\_inclusion":true}. This record does not assert independent reproduction. ### DCL BLEU-4: 10.9. Result ID: 69946e8d3a6c3148491d33cd. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Params":"0.25B","expert\_knowledge\_inclusion":false}. This record does not assert independent reproduction. ### DCL BLEU-4: 10.9. Result ID: 69946e8d3a6c3148491d33ce. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Params":"0.25B","expert\_knowledge\_inclusion":true}. This record does not assert independent reproduction. ### Llava-Med BLEU-4: 1. Result ID: 69946e8d3a6c3148491d33c9. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Params":"7B","expert\_knowledge\_inclusion":false}. This record does not assert independent reproduction. ### Llava-Med BLEU-4: 1. Result ID: 69946e8d3a6c3148491d33ca. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Params":"7B","expert\_knowledge\_inclusion":true}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.