Spatial Understanding on RealWorldQA
66.01RWQA ScoreUniMRG
Evaluation Results
| Method | Links | |
|---|---|---|
| UniMRGType=Unified, # LLM Params=2B2026.01 | 66.01 | |
| InternVL3Type=Und. Only, # LLM Params=2B2026.01 | 65.23 | |
| OpenUniType=Unified, # LLM Params=2B2026.01 | 65.23 | |
| UniLIPType=Unified, # LLM Params=2B2026.01 | 64.18 | |
| Emu3-ChatType=Und. Only, # LLM Params=8B2026.01 | 56.99 | |
| HarmonType=Unified, # LLM Params=1.5B2026.01 | 46.67 | |
| LLaVA-OVType=Und. Only, # LLM Params=0.5B2026.01 | 46.41 | |
| JanusType=Unified, # LLM Params=1.3B2026.01 | 43.66 | |
| Janus-ProType=Unified, # LLM Params=7B2026.01 | 41.83 | |
| Show-oType=Unified, # LLM Params=1.3B2026.01 | 38.17 | |
| Cambrian-34BSize=34B2024.08 | 0.671 | |
| LLaVA-NeXT-Yi-34BSize=34B2024.08 | 0.66 | |
| MiniCPM-Llama3-V 2.5Size=8.5B2024.08 | 0.635 | |
| GPT-4V (2023.11.06)2024.08 | 0.63 | |
| CogVLM2-19B-ChatSize=19B2024.08 | 0.629 | |
| Idefics2Size=8.0B2024.08 | 0.607 | |
| Gemini Pro2024.08 | 0.604 | |
| CogVLM-ChatSize=17B2024.08 | 0.603 | |
| LLaVA-NeXT-Llama3-8BSize=8.4B2024.08 | 0.6 | |
| Cambrian-8BSize=8.3B2024.08 | 0.6 | |
| Phi-3-vision-128k-instructSize=4.2B2024.08 | 0.588 | |
| Bunny-Llama-3-8BSize=8.4B2024.08 | 0.588 | |
| MiniCPM-V 2.0Size=2.8B2024.08 | 0.558 | |
| Yi-VL-34BSize=34B2024.08 | 0.548 | |
| DeepSeek-VL-7BSize=7.3B2024.08 | 0.542 | |
| Yi-VL-6BSize=6.7B2024.08 | 0.535 | |
| MiniCPM-V 1.0Size=2.8B2024.08 | 0.512 | |
| DeepSeek-VL-1.3BSize=1.7B2024.08 | 0.497 | |
| Qwen-VL-ChatSize=9.6B2024.08 | 0.493 | |
| Claude 3 Opus2024.08 | 0.484 |