Text-to-Image Retrieval on MSCOCO (R@1, R@5, R@25, R@50)
62.9Recall@1Gemini Embedding 2
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Gemini Embedding 2Modality=V/A/I/T2026.05 | 62.9 | — | — | — | |
| Voyage-3.5-multimodalModality=V/I/T2026.05 | 58.1 | — | — | — | |
| Amazon Nova MMEModality=V/A/I/T2026.05 | 57.2 | — | — | — | |
| multimodalembedding@001Modality=I/T2026.05 | 53.1 | — | — | — | |
| MobileCLIPv2Seen Data=12B2026.05 | 49.63 | 74.92 | 91.58 | 96.26 | |
| SigLIPv2Seen Data=10B2026.05 | 49.28 | 73.36 | 90.33 | 95.08 | |
| FAST-GOALSeen Data=400M+100k2026.05 | 42.81 | 68.61 | 88.54 | 94.64 | |
| TinyCLIPSeen Data=400M2026.05 | 34.02 | 59.75 | 83.31 | 91.17 |