# Medical Reasoning Segmentation on MR-MedSeg hard case (test) Source: [Medical Reasoning Segmentation on MR-MedSeg hard case (test)]() Collection index: [llms.txt]() Data updated: 2026-03-11T06:44:08.802Z Evaluation of Medical Reasoning Segmentation performance on the test split of the MR-MedSeg hard case dataset, reporting Dice, gIoU, and cIoU scores. 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 | Medical Reasoning Segmentation | | Dataset (including any stated split) | MR-MedSeg hard case (test) | | Metric | Dice | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 15 of 15 | | 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 - [Dice](): Higher is better; unit: PERCENT. - [gIoU](): Higher is better; unit: PERCENT. - [cIoU](): Higher is better; unit: PERCENT. ## Recorded results | Method | Dice | Result date | Report paper | | --- | --- | --- | --- | | MediRound + JCM | 50.3 | Not recorded | 691bfdf46750be83fe5ceb61 | | MediRound | 48.1 | Not recorded | 691bfdf46750be83fe5ceb61 | | MediRound + READ | 47.2 | Not recorded | 691bfdf46750be83fe5ceb61 | | Human-Thinking + MediSee | 45.1 | Not recorded | 691bfdf46750be83fe5ceb61 | | Gemini-2.5-Pro + MediSee | 38.6 | Not recorded | 691bfdf46750be83fe5ceb61 | | Qwen3-VL + MediSee | 37.8 | Not recorded | 691bfdf46750be83fe5ceb61 | | GPT-4o + MediSee | 33.4 | Not recorded | 691bfdf46750be83fe5ceb61 | | SegLLM-13B | 33 | Not recorded | 691bfdf46750be83fe5ceb61 | | Human-Thinking + MedPLIB | 32 | Not recorded | 691bfdf46750be83fe5ceb61 | | SegLLM-7B | 30.4 | Not recorded | 691bfdf46750be83fe5ceb61 | | Human-Thinking + IMIS-Net | 26.4 | Not recorded | 691bfdf46750be83fe5ceb61 | | Qwen3-VL + IMIS-Net | 23.4 | Not recorded | 691bfdf46750be83fe5ceb61 | | Gemini-2.5-Pro + IMIS-Net | 21.9 | Not recorded | 691bfdf46750be83fe5ceb61 | | GPT-4o + IMIS-Net | 17.4 | Not recorded | 691bfdf46750be83fe5ceb61 | | Human-Thinking + BiomedParse | 11.3 | Not recorded | 691bfdf46750be83fe5ceb61 | ### MediRound + JCM Dice: 50.3. Result ID: 69b10f216dbb7075b95acf55. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Independent Model (T)":"true"}. This record does not assert independent reproduction. ### MediRound Dice: 48.1. Result ID: 69b10f216dbb7075b95acf54. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Independent Model (T)":"true"}. This record does not assert independent reproduction. ### MediRound + READ Dice: 47.2. Result ID: 69b10f216dbb7075b95acf53. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Independent Model (T)":"true"}. This record does not assert independent reproduction. ### Human-Thinking + MediSee Dice: 45.1. Result ID: 69b10f216dbb7075b95acf4a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Independent Model (T)":"false"}. This record does not assert independent reproduction. ### Gemini-2.5-Pro + MediSee Dice: 38.6. Result ID: 69b10f216dbb7075b95acf4f. [Reporting paper]() Recorded evaluation configuration: {"Independent Model (T)":"false"}. This record does not assert independent reproduction. ### Qwen3-VL + MediSee Dice: 37.8. Result ID: 69b10f216dbb7075b95acf50. [Reporting paper]() Recorded evaluation configuration: {"Independent Model (T)":"false"}. This record does not assert independent reproduction. ### GPT-4o + MediSee Dice: 33.4. Result ID: 69b10f216dbb7075b95acf4e. [Reporting paper]() Recorded evaluation configuration: {"Independent Model (T)":"false"}. This record does not assert independent reproduction. ### SegLLM-13B Dice: 33. Result ID: 69b10f216dbb7075b95acf52. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Independent Model (T)":"true"}. This record does not assert independent reproduction. ### Human-Thinking + MedPLIB Dice: 32. Result ID: 69b10f216dbb7075b95acf48. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Independent Model (T)":"false"}. This record does not assert independent reproduction. ### SegLLM-7B Dice: 30.4. Result ID: 69b10f216dbb7075b95acf51. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Independent Model (T)":"true"}. This record does not assert independent reproduction. ### Human-Thinking + IMIS-Net Dice: 26.4. Result ID: 69b10f216dbb7075b95acf49. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Independent Model (T)":"false"}. This record does not assert independent reproduction. ### Qwen3-VL + IMIS-Net Dice: 23.4. Result ID: 69b10f216dbb7075b95acf4d. [Reporting paper]() Recorded evaluation configuration: {"Independent Model (T)":"false"}. This record does not assert independent reproduction. ### Gemini-2.5-Pro + IMIS-Net Dice: 21.9. Result ID: 69b10f216dbb7075b95acf4c. [Reporting paper]() Recorded evaluation configuration: {"Independent Model (T)":"false"}. This record does not assert independent reproduction. ### GPT-4o + IMIS-Net Dice: 17.4. Result ID: 69b10f216dbb7075b95acf4b. [Reporting paper]() Recorded evaluation configuration: {"Independent Model (T)":"false"}. This record does not assert independent reproduction. ### Human-Thinking + BiomedParse Dice: 11.3. Result ID: 69b10f216dbb7075b95acf47. [Reporting paper]() Recorded evaluation configuration: {"Independent Model (T)":"false"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.