Medical Image Segmentation Research Automation on 26 blind-test datasets
26Completion CountCamyla
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| CamylaConfiguration=DeepSeek V3.2 (CamylaD)2026.04 | 26 | 14 | 0 | 65.91 | 28.43 | 69.62 | 15.57 | |
| CamylaConfiguration=Claude Sonnet 4.6 (CamylaS)2026.04 | 26 | 18 | 0 | 65.16 | 29.21 | 69.62 | 16.01 | |
| autoresearchConfiguration=Claude Code + Opus 4.62026.04 | 18 | 10 | 4 | 63.14 | 34.72 | 67.25 | 20.83 | |
| AI ScientistConfiguration=Default workflow2026.04 | 18 | 5 | — | 60.24 | 40.55 | 63.18 | 26.31 | |
| autoresearchConfiguration=Claude Code + GLM 4.72026.04 | 16 | 10 | 12 | 62.21 | 37.06 | 65.89 | 23.04 | |
| autoresearchConfiguration=Claude Code + MiniMax 2.52026.04 | 15 | 12 | 14 | 62.58 | 36.41 | 66.43 | 22.15 | |
| autoresearchConfiguration=Codex + GPT-5.4 (xhigh)2026.04 | 12 | 8 | 10 | 61.35 | 38.92 | 64.71 | 24.67 | |
| AI Scientist-v2Configuration=BFTS workflow2026.04 | 12 | 3 | — | 58.73 | 43.18 | 61.52 | 28.94 |