Multi-modal Visual Capability on MMStar
63.9ScoreGPT-4o-0513
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-4o-05132024.09 | 63.9 | |
| Claude3.5-Sonnet2024.09 | 62.2 | |
| OneVision2024.09 | 61.9 | |
| Defender Iter. 3Training Strategy=Iterative Co-evolution2026.01 | 61.53 | |
| InternVL2-8B2024.09 | 61.5 | |
| Defender Iter. 2Training Strategy=Iterative Co-evolution2026.01 | 61.4 | |
| Defender Iter. 1Training Strategy=Iterative Co-evolution2026.01 | 61.27 | |
| Liu et al. (Insert)Training Strategy=Finite Augmentation Datasets2026.01 | 61.2 | |
| POINTS-7BLLM=Qwen-2.5-7B2024.09 | 61 | |
| Ablation (Clean Data)Training Strategy=Ablation2026.01 | 60.93 | |
| Yang et al.Training Strategy=Finite Augmentation Datasets2026.01 | 60.87 | |
| Liu et al. (All)Training Strategy=Finite Augmentation Datasets2026.01 | 60.8 | |
| Liu et al. (Add)Training Strategy=Finite Augmentation Datasets2026.01 | 60.4 | |
| Base (M_def^(0))Training Strategy=Base2026.01 | 60.33 | |
| IXC-2.52024.09 | 59.9 | |
| Gemini-1.5-Pro2024.09 | 59.1 | |
| Ovis1.5-LLAMA3-8B2024.09 | 57.3 | |
| POINTS-9BLLM=Yi-1.5-9B2024.09 | 56.4 | |
| Idefics3-LLAMA3-8B2024.09 | 55 | |
| Cambrian-34B2024.09 | 54.2 | |
| Emu2-ChatType=Understanding-only, LLM Params=33B2025.03 | 40.7 | |
| Qwen-VLType=Understanding-only, LLM Params=7B2025.03 | 32.5 | |
| Chameleon-7bType=Unified (Understanding & Generation), LLM Params=7B2025.03 | 31.1 | |
| WeGenType=Unified (Understanding & Generation), LLM Params=7B2025.03 | 27.5 | |
| LLaVAType=Understanding-only, LLM Params=7B2025.03 | 27.1 | |
| Kosmos-2Type=Understanding-only, LLM Params=2B2025.03 | 24.9 | |
| IdeficsType=Understanding-only, LLM Params=9B2025.03 | 21.6 | |
| Kosmos-GType=Unified (Understanding & Generation), LLM Params=1.9B2025.03 | 18.4 | |
| MiniGPT4Type=Understanding-only, LLM Params=7B2025.03 | 16.3 |