Rare-disease diagnosis on CROSSRARE-BENCH Overall 1.0 (test)
73AccuracyMedLatentDx
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MedLatentDxModel=Llama-3.2-3B, Collaboration Category=Latent communication2026.06 | 73 | 68 | 77 | |
| Struct. RAGModel=Qwen3-4B, Collaboration Category=Text / structured sharing2026.06 | 64 | 52 | 71 | |
| MedLatentDxModel=Qwen3-4B, Collaboration Category=Latent communication2026.06 | 64 | 61 | 72 | |
| Global RAGModel=Qwen3-4B, Collaboration Category=Text / structured sharing2026.06 | 62 | 53 | 70 | |
| Dist. RAGModel=Qwen3-4B, Collaboration Category=Text / structured sharing2026.06 | 61 | 53 | 72 | |
| LatentMASModel=Qwen3-4B, Collaboration Category=Latent communication2026.06 | 60 | 49 | 71 | |
| Raw KVModel=Qwen3-4B, Collaboration Category=Latent communication2026.06 | 59 | 48 | 72 | |
| Single RAGModel=Qwen3-4B, Collaboration Category=Non-collaborative2026.06 | 57 | 47 | 66 | |
| TextMASModel=Qwen3-4B, Collaboration Category=Text / structured sharing2026.06 | 54 | 44 | 69 | |
| Global RAGModel=Llama-3.2-3B, Collaboration Category=Text / structured sharing2026.06 | 47 | 42 | 62 | |
| Dist. RAGModel=Llama-3.2-3B, Collaboration Category=Text / structured sharing2026.06 | 45 | 40 | 63 | |
| Struct. RAGModel=Llama-3.2-3B, Collaboration Category=Text / structured sharing2026.06 | 42 | 34 | 67 | |
| Single RAGModel=Llama-3.2-3B, Collaboration Category=Non-collaborative2026.06 | 39 | 34 | 57 | |
| Raw KVModel=Llama-3.2-3B, Collaboration Category=Latent communication2026.06 | 35 | 32 | 58 | |
| LatentMASModel=Llama-3.2-3B, Collaboration Category=Latent communication2026.06 | 33 | 31 | 59 | |
| TextMASModel=Llama-3.2-3B, Collaboration Category=Text / structured sharing2026.06 | 32 | 25 | 47 | |
| Zero-shotModel=Qwen3-4B, Collaboration Category=Non-collaborative2026.06 | 3 | 2 | 24 | |
| Zero-shotModel=Llama-3.2-3B, Collaboration Category=Non-collaborative2026.06 | 1 | 0 | 11 |