Our new X account is live! Follow @wizwand_team for updates

Vision-Language Navigation on EmbodiedBench Navigation latest (test)

86.7Success Rate

GPT4.1+VP

Updated 4d ago

Evaluation Results

Method	Links
GPT4.1+VP 2025.12		86.7
VP+SRGPO 2025.12		72.3
Claude-3.5-Sonnet 2025.12		66.7
InternVL3-78B 2025.12		66.7
GPT4.1 2025.12		65
Gemini-2.0-flash 2025.12		63.3
Ovis2-34B 2025.12		63.3
VP+GiGPO (w. VPR) 2025.12		57.2
GPT-4o 2025.12		55
gemma-3-27b-it 2025.12		53.3
Claude-3.7-Sonnet 2025.12		50
Qwen-VL-Max 2025.12		50
Llama-3.2-11B-Vision-Ins 2025.12		48.3
Qwen2.5-VL-72B-Ins 2025.12		46.7
VP+GRPO 2025.12		40.8
InternVL3-8B 2025.12		38.3
VP+SFT 2025.12		36.7
GPT-4o-mini 2025.12		31.7
VP+GiGPO (vanilla) 2025.12		29.4
Gemini-1.5-Pro 2025.12		23.3
Llama-3.2-90B-Vision-Ins 2025.12		23.3
Qwen2.5-VL-7B-Ins 2025.12		20
Qwen2.5-VL-3B-Ins 2025.12		16.7