| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Perturbation prediction | PBMC | DE Overlap81 | 22 | |
| Single-cell Denoising | PBMC OpenProblems benchmark | Mean Score0.711 | 11 | |
| Single-Cell Gene Regulatory Network Inference | PBMC human | AUPRC0.244 | 10 | |
| On-domain cell identity prediction | PBMC10K | Macro F1 Score96 | 10 | |
| Denoising | PBMC (held-out) | Score0.71 | 9 | |
| Clustering | PBMC 3k | ARI92 | 9 | |
| Gene Representation Learning | PBMC (test) | Average Score92 | 9 | |
| Gene embedding | PBMC | Average Rank2.33 | 9 | |
| Cell type clustering | PBMC | ARI0.64 | 8 | |
| clustering | PBMC | ARI0.743 | 7 | |
| Representation Learning | PBMC3k | Rec Score88 | 7 | |
| scRNA-seq integration | pbmc | Batch Correction88.6 | 7 | |
| Symbolic Density Estimation | PBMC gene 4046 | MSE0.1263 | 6 | |
| Biological process inference | pbmc3k | Mean LLM Confidence Score90.21 | 6 | |
| Gene rank reconstruction | PBMC12k | L-Dist6 | 6 | |
| Denoising | PBMC10k | SCC0.916 | 6 | |
| Gene Regulatory Network Inference | PBMC | p-value0.01 | 5 | |
| Single-cell Differential Abundance Estimation | PBMC semi-synthetic 68k cells | Correlation (rho) for AUC1 | 5 | |
| Batch effect correction | PBMC | Biological Conservation Score0.75 | 5 | |
| Multi-batch integration | PBMC12k | Batch Score0.9755 | 5 | |
| Clustering | PBMC | Homogeneity79.8 | 5 | |
| Dual condition generation | PBMC | WD5.37 | 4 | |
| Gene level performance | PBMC3k | MSE0.07 | 4 | |
| Single conditional generation | PBMC3K | WD16.94 | 4 | |
| Gene level performance evaluation | PBMC | MSE0.02 | 4 |