Reasoning Segmentation on ReasonSeg short (test)
59.4cIoUSAM 3 Agent
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SAM 3 AgentVersion=Llama4 Maverick, Training-RES=false, Training-ReasonSeg=false2025.12 | 59.4 | 66.8 | |
| HRSegVersion=LLaVA1.6 7B, Training-RES=true, Training-ReasonSeg=true2025.12 | 58.9 | 59.2 | |
| SAM 3 AgentVersion=Qwen2.5-VL 72B, Training-RES=false, Training-ReasonSeg=false2025.12 | 55.7 | 70.3 | |
| HRSegVersion=LLaVA1.1 7B, Training-RES=true, Training-ReasonSeg=true2025.12 | 54.5 | 54.4 | |
| RSAgentModel Version=Qwen2.5-VL, Params=7B, RES Training=true, ReasonSeg Training=false2025.12 | 54.4 | 60.3 | |
| LISAVersion=LLaVA1.5 13B, Training-RES=true, Training-ReasonSeg=true2025.12 | 50.6 | 55.4 | |
| RSVPVersion=GPT-4o, Training-RES=false, Training-ReasonSeg=false2025.12 | 50.4 | 55.4 | |
| RSVPModel Version=GPT-4o, RES Training=false, ReasonSeg Training=false2025.12 | 50.4 | 55.4 | |
| LISAVersion=LLaVA1.5 13B, Training-RES=false, Training-ReasonSeg=true2025.12 | 50 | 50.8 | |
| READVersion=LLaVA1.5 7B, Training-RES=true, Training-ReasonSeg=true2025.12 | 49.5 | 52.6 | |
| LISAVersion=LLaVA1.5 7B, Training-RES=true, Training-ReasonSeg=false2025.12 | 48.8 | 48.3 | |
| LISAModel Version=LLaVA1.5, Params=7B, RES Training=true, ReasonSeg Training=false2025.12 | 48.8 | 48.3 | |
| Ref-DiffVersion=LLaVA1.5 13B, Training-RES=false, Training-ReasonSeg=false2025.12 | 48.7 | — | |
| X-SAMVersion=Phi3 3.8B, Training-RES=true, Training-ReasonSeg=true2025.12 | 48.1 | 47.7 | |
| X-SAMModel Version=Phi-3, Params=3.8B, RES Training=true, ReasonSeg Training=true2025.12 | 48.1 | 47.7 | |
| Ref-DiffVersion=LLaVA1.5 7B, Training-RES=false, Training-ReasonSeg=false2025.12 | 48 | — | |
| RSAgent-singleModel Version=Qwen2.5-VL, Params=7B, RES Training=true, ReasonSeg Training=false2025.12 | 47.7 | 54.9 | |
| EVOL-SAM3Version=Qwen2.5-VL 3B, Training-RES=false, Training-ReasonSeg=false2025.12 | 47.3 | 62 | |
| EVOL-SAM3Version=Qwen2.5-VL 7B, Training-RES=false, Training-ReasonSeg=false2025.12 | 46.8 | 67 | |
| LISAVersion=LLaVA1.5 7B, Training-RES=true, Training-ReasonSeg=true2025.12 | 46.3 | 48.3 | |
| LISAVersion=Llama2 13B, Training-RES=true, Training-ReasonSeg=true2025.12 | 45.8 | 43.9 | |
| RSVPVersion=Qwen2-VL 7B, Training-RES=false, Training-ReasonSeg=false2025.12 | 44.3 | 48.5 | |
| SAM 3 AgentVersion=Qwen2.5-VL 7B, Training-RES=false, Training-ReasonSeg=false2025.12 | 43.5 | 59.4 | |
| LISAVersion=LLaVA 13B, Training-RES=true, Training-ReasonSeg=false2025.12 | 43.3 | 39.9 | |
| LISAVersion=LLaVA 13B, Training-RES=true, Training-ReasonSeg=true2025.12 | 42 | 44.3 | |
| RSVPVersion=LLaVA1.6 7B, Training-RES=false, Training-ReasonSeg=false2025.12 | 42 | 47.9 | |
| LISAModel Version=LLaVA1.5, Params=7B, RES Training=true, ReasonSeg Training=true2025.12 | 42 | 44.3 | |
| RSVPModel Version=LLaVA1.6, Params=7B, RES Training=false, ReasonSeg Training=false2025.12 | 42 | 47.9 | |
| SAM3-AgentModel Version=Qwen2.5-VL, Params=7B, RES Training=false, ReasonSeg Training=false2025.12 | 41.8 | 59.1 | |
| LISAVersion=LLaVA 7B, Training-RES=true, Training-ReasonSeg=true2025.12 | 40.6 | 40.6 | |
| LLM-SegVersion=LLaVA1.1 7B, Training-RES=true, Training-ReasonSeg=true2025.12 | 40.2 | 41.1 | |
| RSVPVersion=Gemini1.5-Flash, Training-RES=false, Training-ReasonSeg=false2025.12 | 40.2 | 47.3 | |
| LISAVersion=LLaVA 7B, Training-RES=true, Training-ReasonSeg=false2025.12 | 34.4 | 37.6 | |
| OVSegVersion=–, Training-RES=false, Training-ReasonSeg=false2025.12 | 15.5 | 18 | |
| OVSegRES Training=false, ReasonSeg Training=false2025.12 | 15.5 | 18 | |
| SEEMVersion=–, Training-RES=false, Training-ReasonSeg=false2025.12 | 11.5 | 20.1 | |
| SEEMRES Training=false, ReasonSeg Training=false2025.12 | 11.5 | 20.1 | |
| Grounded SAMVersion=–, Training-RES=false, Training-ReasonSeg=false2025.12 | 10.8 | 17.8 | |
| Grounded-SAMRES Training=false, ReasonSeg Training=false2025.12 | 10.8 | 17.8 | |
| Gemini SegVersion=Gemini2.5 Flash, Training-RES=?, Training-ReasonSeg=?2025.12 | 8 | 16.5 |