Spatial Reasoning on BLINK
69.1ScoreQ3VL 8B-Ins
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Q3VL 8B-InsRL=N/A, Initial Model=N/A2026.04 | 69.1 | — | |
| Vero Q3I-8BRL=Qwen3VL, Initial Model=8B Inst2026.04 | 68.7 | — | |
| GeoSenseFT-Data=940K, Fine-tuned on identical training data=false2026.03 | 68.5 | — | |
| Vero Q3T-8BRL=Qwen3VL, Initial Model=8B Think2026.04 | 66.3 | — | |
| Q3VL 8B-ThkRL=N/A, Initial Model=N/A2026.04 | 64.7 | — | |
| MiMoVL 7B-RLRL=MiMoVL, Initial Model=7B-SFT2026.04 | 64.5 | — | |
| Qwen2.5-VL-3B*FT-Data=940k, Fine-tuned on identical training data=true2026.03 | 63.1 | — | |
| VG-LLM*FT-Data=940k, Fine-tuned on identical training data=true2026.03 | 62.8 | — | |
| Vero Mi-7BRL=MiMoVL, Initial Model=7B-SFT2026.04 | 62.3 | — | |
| Vero Q25-7BRL=Qwen25VL, Initial Model=7B Inst2026.04 | 59.9 | — | |
| SpatialImaginer2026.04 | 59.8 | — | |
| GPT-5 NanoRL=N/A, Initial Model=N/A2026.04 | 59.3 | — | |
| Q25VL 7B-InsRL=N/A, Initial Model=N/A2026.04 | 56.4 | — | |
| Mo2-O 7BRL=SFT Only, Initial Model=N/A2026.04 | 56.4 | — | |
| Qwen2.5-VL-7BFT-Data=-, Fine-tuned on identical training data=false2026.03 | 55.9 | — | |
| VLM-3R-7B2026.04 | 52.3 | — | |
| ViLASR-7BFT-Data=73K, Fine-tuned on identical training data=false2026.03 | 51.4 | — | |
| ViLaSR-7B2026.04 | 51.4 | — | |
| SpaceR-sft-7BFT-Data=151K, Fine-tuned on identical training data=false2026.03 | 49.6 | — | |
| SpaceR-7B2026.04 | 49.6 | — | |
| VG-LLMFT-Data=385k, Fine-tuned on identical training data=false2026.03 | 48.9 | — | |
| SpatialLadder-3BFT-Data=26K, Fine-tuned on identical training data=false2026.03 | 43 | — | |
| SpatialLadder-3B2026.04 | 43 | — | |
| Spatial-MLLM-4B2026.04 | 40.5 | — | |
| Cambrian-S-7B2026.04 | 37.9 | — | |
| Cambrain-S-3BFT-Data=10M, Fine-tuned on identical training data=false2026.03 | 37.7 | — | |
| MindCube-3B2026.04 | 35.1 | — | |
| Qwen2.5-VL-3BFT-Data=-, Fine-tuned on identical training data=false2026.03 | 33.1 | — | |
| InternVL3-2BFT-Data=-, Fine-tuned on identical training data=false2026.03 | 30 | — | |
| Claude Sonnet 4.5Model Category=Proprietary Models2025.12 | — | 78.23 | |
| Gemini-ER 1.5Model Category=Proprietary Models2025.12 | — | 69.23 | |
| GPT-4oModel Category=Proprietary Models2025.12 | — | 63.71 | |
| GPT-5Model Category=Proprietary Models2025.12 | — | 66.13 | |
| LLaVA-NeXT-8BModel Category=General Open-Source Models2025.12 | — | 53.23 | |
| Molmo-7BModel Category=Spatial VLMs2025.12 | — | 54.03 | |
| Qwen2.5-VL-32BModel Category=General Open-Source Models2025.12 | — | 70.16 | |
| Qwen2.5-VL-3BModel Category=General Open-Source Models2025.12 | — | 70.98 | |
| Qwen2.5-VL-3B-Tool-freeModel Category=Tool-free Fine-tuning, Fine-tuning Protocol=SFT2025.12 | — | 80.65 | |
| Qwen2.5-VL-3B-Tool-freeModel Category=Tool-free Fine-tuning, Fine-tuning Protocol=RL2025.12 | — | 80.65 | |
| RoboBrain2.0-7BModel Category=Spatial VLMs2025.12 | — | 84.68 | |
| RoboPoint-13BModel Category=Spatial VLMs2025.12 | — | 54.84 | |
| RoboRefer-8B-SFTModel Category=Spatial VLMs2025.12 | — | 88.71 | |
| SpaceLLaVA-13BModel Category=Spatial VLMs2025.12 | — | 51.61 | |
| SpaceTools-3BModel Category=Ours2025.12 | — | 90.32 |