Text Question Answering on HELMET
78.4AccuracyLensVLM
Evaluation Results
| Method | Links | |
|---|---|---|
| LensVLMICR=5×, ECR=4.3×2026.05 | 78.4 | |
| TextICR=1×, ECR=1.0×2026.05 | 76.2 | |
| BGE-M3ICR=5×, ECR=4.5×2026.05 | 73.8 | |
| Jina-v4ICR=5×, ECR=4.5×2026.05 | 73.7 | |
| Jina-v4-MMICR=5×, ECR=4.6×2026.05 | 73 | |
| Qwen3-EmbICR=5×, ECR=4.5×2026.05 | 72.7 | |
| LensVLMSFTICR=5×, ECR=4.6×2026.05 | 72.7 | |
| ColPaliICR=5×, ECR=4.7×2026.05 | 71.5 | |
| LensVLMICR=10×, ECR=7.4×2026.05 | 71.5 | |
| BM25ICR=5×, ECR=4.5×2026.05 | 70.3 | |
| Qwen3-EmbICR=10×, ECR=8.3×2026.05 | 70 | |
| Jina-v4ICR=10×, ECR=8.3×2026.05 | 69.7 | |
| Jina-v4-MMICR=15×, ECR=11.9×2026.05 | 69 | |
| Jina-v4ICR=15×, ECR=11.9×2026.05 | 68.8 | |
| BGE-M3ICR=10×, ECR=8.3×2026.05 | 68.3 | |
| BGE-M3ICR=15×, ECR=11.9×2026.05 | 67.8 | |
| Qwen3-EmbICR=15×, ECR=11.9×2026.05 | 67.5 | |
| BM25ICR=10×, ECR=8.3×2026.05 | 65.5 | |
| GlyphICR=5×, ECR=4.5×2026.05 | 63.5 | |
| BM25ICR=15×, ECR=11.9×2026.05 | 63.5 | |
| LensVLMSFTICR=10×, ECR=8.1×2026.05 | 63.2 | |
| Jina-v4-MMICR=10×, ECR=8.3×2026.05 | 63 | |
| ColPaliICR=10×, ECR=8.3×2026.05 | 60 | |
| LLMLingua-2ICR=5×, ECR=5.5×2026.05 | 58.7 | |
| LensVLMICR=15×, ECR=10.1×2026.05 | 57.2 | |
| ColPaliICR=15×, ECR=11.9×2026.05 | 57 | |
| LensVLMSFTICR=15×, ECR=10.7×2026.05 | 51.7 | |
| LLMLingua-2ICR=10×, ECR=11.3×2026.05 | 50.8 | |
| GlyphICR=10×, ECR=8.6×2026.05 | 50.5 | |
| LLMLingua-2ICR=15×, ECR=15.7×2026.05 | 47.2 | |
| GlyphICR=15×, ECR=10.8×2026.05 | 46 | |
| Base + ExpandICR=5×, ECR=4.5×2026.05 | 35 | |
| Comp. ImageICR=5×, ECR=5×2026.05 | 33.1 | |
| Base + ExpandICR=10×, ECR=8.0×2026.05 | 30.5 | |
| Comp. ImageICR=15×, ECR=15×2026.05 | 24.5 | |
| Base + ExpandICR=15×, ECR=11.0×2026.05 | 24.5 | |
| Comp. ImageICR=10×, ECR=10×2026.05 | 22.5 |