Multimodal Mathematical Reasoning on MathVista (MathVista score, Avg-R)
76.4MathVista ScoreSPOT-E
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SPOT-EBase Model=Qwen3-VL-32B2026.06 | 76.4 | — | |
| SPOT-EBase Model=Qwen2.5-VL-32B2026.06 | 75.3 | — | |
| Qwen3-VL-32BBase Model=Qwen3-VL-32B2026.06 | 75.1 | — | |
| SPOT-EBase Model=InternVL2.5-26B2026.06 | 74.3 | — | |
| Qwen2.5-VL-32BBase Model=Qwen2.5-VL-32B2026.06 | 73.8 | — | |
| InternVL2.5-26BBase Model=InternVL2.5-26B2026.06 | 72.5 | — | |
| GAP (LH+PCA+DA)Latent Head (LH)=true, Difficulty-Aware supervision (DA)=true, PCA alignment (PCA)=true2026.05 | 71 | 53.97 | |
| GAP (LH+PCA)Latent Head (LH)=true, PCA alignment (PCA)=true2026.05 | 70.2 | 52.48 | |
| GAP (LH+DA, no PCA)Latent Head (LH)=true, Difficulty-Aware supervision (DA)=true, PCA alignment (PCA)=false2026.05 | 68.9 | 50.05 | |
| GAP (LH)Latent Head (LH)=true2026.05 | 68.6 | 49.15 | |
| Qwen2.5-VL 7BParameters=7B2026.05 | 68.5 | 52.62 | |
| Dense Cap SFT2026.05 | 67.7 | 47.24 | |
| LVR2026.05 | 67 | 47.66 | |
| Monet-7BParameters=7B2026.05 | 61.3 | 47.99 | |
| SPOT-EBase Model=LLaVA-OV-72B2026.06 | 60.3 | — | |
| LLaVA-OV-72BBase Model=LLaVA-OV-72B2026.06 | 58 | — |