# Medical Diagnostic Reasoning on MedCaseReasoning official 897-case (test) Source: [Medical Diagnostic Reasoning on MedCaseReasoning official 897-case (test)]() Collection index: [llms.txt]() Data updated: 2026-04-14T11:22:35.748Z Evaluation of model performance on the official 897-case test split of the MedCaseReasoning dataset, focusing on Medical Diagnostic Reasoning capabilities, measured by Accuracy and TrustScore. 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 Diagnostic Reasoning | | Dataset (including any stated split) | MedCaseReasoning official 897-case (test) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 36 of 36 | | 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 - [Accuracy](): Higher is better; unit: PERCENT. - [Trust Score](): Higher is better; unit: SCALAR. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | GRPO | 32.83 | Not recorded | 69ddca6c7d0172140ec79edb | | GRPO | 32.81 | Not recorded | 69ddca6c7d0172140ec79edb | | CGCL | 32.67 | Not recorded | 69ddca6c7d0172140ec79edb | | CGCL | 32.63 | Not recorded | 69ddca6c7d0172140ec79edb | | DPO | 32.29 | Not recorded | 69ddca6c7d0172140ec79edb | | DPO | 32.24 | Not recorded | 69ddca6c7d0172140ec79edb | | SFT-CoT | 32.14 | Not recorded | 69ddca6c7d0172140ec79edb | | SFT-CoT | 32.08 | Not recorded | 69ddca6c7d0172140ec79edb | | SFT-GT | 31.79 | Not recorded | 69ddca6c7d0172140ec79edb | | SFT-GT | 31.72 | Not recorded | 69ddca6c7d0172140ec79edb | | SFT-CGCL(Stage3) | 30.12 | Not recorded | 69ddca6c7d0172140ec79edb | | SFT-CGCL(Stage3) | 30.07 | Not recorded | 69ddca6c7d0172140ec79edb | | CGCL | 28.95 | Not recorded | 69ddca6c7d0172140ec79edb | | GRPO | 28.95 | Not recorded | 69ddca6c7d0172140ec79edb | | GRPO | 28.9 | Not recorded | 69ddca6c7d0172140ec79edb | | CGCL | 28.9 | Not recorded | 69ddca6c7d0172140ec79edb | | DPO | 28.63 | Not recorded | 69ddca6c7d0172140ec79edb | | DPO | 28.59 | Not recorded | 69ddca6c7d0172140ec79edb | | SFT-CoT | 28.33 | Not recorded | 69ddca6c7d0172140ec79edb | | SFT-CoT | 28.27 | Not recorded | 69ddca6c7d0172140ec79edb | | SFT-GT | 28.16 | Not recorded | 69ddca6c7d0172140ec79edb | | SFT-GT | 28.12 | Not recorded | 69ddca6c7d0172140ec79edb | | SFT-CGCL(Stage3) | 26.64 | Not recorded | 69ddca6c7d0172140ec79edb | | SFT-CGCL(Stage3) | 26.58 | Not recorded | 69ddca6c7d0172140ec79edb | | Trust-Think & Answer | 26.38 | Not recorded | 69ddca6c7d0172140ec79edb | | Trust-Think & Answer | 26.31 | Not recorded | 69ddca6c7d0172140ec79edb | | Think & Answer | 25.73 | Not recorded | 69ddca6c7d0172140ec79edb | | Think & Answer | 25.68 | Not recorded | 69ddca6c7d0172140ec79edb | | Think & Answer | 23.51 | Not recorded | 69ddca6c7d0172140ec79edb | | Think & Answer | 23.47 | Not recorded | 69ddca6c7d0172140ec79edb | | Trust-Think & Answer | 23.43 | Not recorded | 69ddca6c7d0172140ec79edb | | Trust-Think & Answer | 23.42 | Not recorded | 69ddca6c7d0172140ec79edb | | Vanilla | 22.46 | Not recorded | 69ddca6c7d0172140ec79edb | | Vanilla | 22.37 | Not recorded | 69ddca6c7d0172140ec79edb | | Vanilla | 20.61 | Not recorded | 69ddca6c7d0172140ec79edb | | Vanilla | 20.57 | Not recorded | 69ddca6c7d0172140ec79edb | ### GRPO Accuracy: 32.83. Result ID: 69de237047f34ce924c6957c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Category":"RL","Backbone":"LLaMA3.1-8B-Instruct"}. This record does not assert independent reproduction. ### GRPO Accuracy: 32.81. Result ID: 69de237047f34ce924c6956a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Category":"RL","Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### CGCL Accuracy: 32.67. Result ID: 69de237047f34ce924c6957d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"Ours","Backbone":"LLaMA3.1-8B-Instruct"}. This record does not assert independent reproduction. ### CGCL Accuracy: 32.63. Result ID: 69de237047f34ce924c6956b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"Ours","Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### DPO Accuracy: 32.29. Result ID: 69de237047f34ce924c6957b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"RL","Backbone":"LLaMA3.1-8B-Instruct"}. This record does not assert independent reproduction. ### DPO Accuracy: 32.24. Result ID: 69de237047f34ce924c69569. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"RL","Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### SFT-CoT Accuracy: 32.14. Result ID: 69de237047f34ce924c69579. [Reporting paper]() Recorded evaluation configuration: {"Category":"SFT","Backbone":"LLaMA3.1-8B-Instruct"}. This record does not assert independent reproduction. ### SFT-CoT Accuracy: 32.08. Result ID: 69de237047f34ce924c69567. [Reporting paper]() Recorded evaluation configuration: {"Category":"SFT","Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### SFT-GT Accuracy: 31.79. Result ID: 69de237047f34ce924c69578. [Reporting paper]() Recorded evaluation configuration: {"Category":"SFT","Backbone":"LLaMA3.1-8B-Instruct"}. This record does not assert independent reproduction. ### SFT-GT Accuracy: 31.72. Result ID: 69de237047f34ce924c69566. [Reporting paper]() Recorded evaluation configuration: {"Category":"SFT","Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### SFT-CGCL(Stage3) Accuracy: 30.12. Result ID: 69de237047f34ce924c69568. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"SFT","Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### SFT-CGCL(Stage3) Accuracy: 30.07. Result ID: 69de237047f34ce924c6957a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"SFT","Backbone":"LLaMA3.1-8B-Instruct"}. This record does not assert independent reproduction. ### CGCL Accuracy: 28.95. Result ID: 69de237047f34ce924c69574. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"Ours","Backbone":"Qwen2.5-3B-Instruct"}. This record does not assert independent reproduction. ### GRPO Accuracy: 28.95. Result ID: 69de237047f34ce924c69585. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Category":"RL","Backbone":"LLaMA3.2-3B-Instruct"}. This record does not assert independent reproduction. ### GRPO Accuracy: 28.9. Result ID: 69de237047f34ce924c69573. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Category":"RL","Backbone":"Qwen2.5-3B-Instruct"}. This record does not assert independent reproduction. ### CGCL Accuracy: 28.9. Result ID: 69de237047f34ce924c69586. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"Ours","Backbone":"LLaMA3.2-3B-Instruct"}. This record does not assert independent reproduction. ### DPO Accuracy: 28.63. Result ID: 69de237047f34ce924c69584. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"RL","Backbone":"LLaMA3.2-3B-Instruct"}. This record does not assert independent reproduction. ### DPO Accuracy: 28.59. Result ID: 69de237047f34ce924c69572. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"RL","Backbone":"Qwen2.5-3B-Instruct"}. This record does not assert independent reproduction. ### SFT-CoT Accuracy: 28.33. Result ID: 69de237047f34ce924c69582. [Reporting paper]() Recorded evaluation configuration: {"Category":"SFT","Backbone":"LLaMA3.2-3B-Instruct"}. This record does not assert independent reproduction. ### SFT-CoT Accuracy: 28.27. Result ID: 69de237047f34ce924c69570. [Reporting paper]() Recorded evaluation configuration: {"Category":"SFT","Backbone":"Qwen2.5-3B-Instruct"}. This record does not assert independent reproduction. ### SFT-GT Accuracy: 28.16. Result ID: 69de237047f34ce924c6956f. [Reporting paper]() Recorded evaluation configuration: {"Category":"SFT","Backbone":"Qwen2.5-3B-Instruct"}. This record does not assert independent reproduction. ### SFT-GT Accuracy: 28.12. Result ID: 69de237047f34ce924c69581. [Reporting paper]() Recorded evaluation configuration: {"Category":"SFT","Backbone":"LLaMA3.2-3B-Instruct"}. This record does not assert independent reproduction. ### SFT-CGCL(Stage3) Accuracy: 26.64. Result ID: 69de237047f34ce924c69583. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"SFT","Backbone":"LLaMA3.2-3B-Instruct"}. This record does not assert independent reproduction. ### SFT-CGCL(Stage3) Accuracy: 26.58. Result ID: 69de237047f34ce924c69571. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Category":"SFT","Backbone":"Qwen2.5-3B-Instruct"}. This record does not assert independent reproduction. ### Trust-Think & Answer Accuracy: 26.38. Result ID: 69de237047f34ce924c69577. [Reporting paper]() Recorded evaluation configuration: {"Category":"Prompt","Backbone":"LLaMA3.1-8B-Instruct"}. This record does not assert independent reproduction. ### Trust-Think & Answer Accuracy: 26.31. Result ID: 69de237047f34ce924c69565. [Reporting paper]() Recorded evaluation configuration: {"Category":"Prompt","Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### Think & Answer Accuracy: 25.73. Result ID: 69de237047f34ce924c69576. [Reporting paper]() Recorded evaluation configuration: {"Category":"Prompt","Backbone":"LLaMA3.1-8B-Instruct"}. This record does not assert independent reproduction. ### Think & Answer Accuracy: 25.68. Result ID: 69de237047f34ce924c69564. [Reporting paper]() Recorded evaluation configuration: {"Category":"Prompt","Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### Think & Answer Accuracy: 23.51. Result ID: 69de237047f34ce924c6957f. [Reporting paper]() Recorded evaluation configuration: {"Category":"Prompt","Backbone":"LLaMA3.2-3B-Instruct"}. This record does not assert independent reproduction. ### Think & Answer Accuracy: 23.47. Result ID: 69de237047f34ce924c6956d. [Reporting paper]() Recorded evaluation configuration: {"Category":"Prompt","Backbone":"Qwen2.5-3B-Instruct"}. This record does not assert independent reproduction. ### Trust-Think & Answer Accuracy: 23.43. Result ID: 69de237047f34ce924c69580. [Reporting paper]() Recorded evaluation configuration: {"Category":"Prompt","Backbone":"LLaMA3.2-3B-Instruct"}. This record does not assert independent reproduction. ### Trust-Think & Answer Accuracy: 23.42. Result ID: 69de237047f34ce924c6956e. [Reporting paper]() Recorded evaluation configuration: {"Category":"Prompt","Backbone":"Qwen2.5-3B-Instruct"}. This record does not assert independent reproduction. ### Vanilla Accuracy: 22.46. Result ID: 69de237047f34ce924c69575. [Reporting paper]() Recorded evaluation configuration: {"Category":"Prompt","Backbone":"LLaMA3.1-8B-Instruct"}. This record does not assert independent reproduction. ### Vanilla Accuracy: 22.37. Result ID: 69de237047f34ce924c69563. [Reporting paper]() Recorded evaluation configuration: {"Category":"Prompt","Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### Vanilla Accuracy: 20.61. Result ID: 69de237047f34ce924c6956c. [Reporting paper]() Recorded evaluation configuration: {"Category":"Prompt","Backbone":"Qwen2.5-3B-Instruct"}. This record does not assert independent reproduction. ### Vanilla Accuracy: 20.57. Result ID: 69de237047f34ce924c6957e. [Reporting paper]() Recorded evaluation configuration: {"Category":"Prompt","Backbone":"LLaMA3.2-3B-Instruct"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.