List your company

Grounded Task Planning on GroundedPlanBench Implicit Instructions, Long Horizon

17.9TSRGemini-3-Flash
-0.7164.1178.9513.783Mar 13, 2026

Evaluation results

MethodLinks
Gemini-3-FlashVLM Category=Proprietary VLMs, Planning Approach=End-to-End Spatially Grounded Planning2026.0317.955.9
Qwen3-VL-32B + V2GPVLM Category=Open-Source VLMs, Planning Approach=End-to-End Spatially Grounded Planning2026.0312.938
Qwen3-VL-4B + V2GPVLM Category=Open-Source VLMs, Planning Approach=End-to-End Spatially Grounded Planning2026.037.931.9
Gemini-2.5-FlashVLM Category=Proprietary VLMs, Planning Approach=End-to-End Spatially Grounded Planning2026.03522.8
GPT 5.2 + ER1VLM Category=Proprietary VLMs, Planning Approach=Decoupled Task Planning + Spatial Grounding2026.033.638.4
Qwen3-VL-32B + ER1VLM Category=Open-Source VLMs, Planning Approach=Decoupled Task Planning + Spatial Grounding2026.032.935.6
Qwen3-VL-32BVLM Category=Open-Source VLMs, Planning Approach=End-to-End Spatially Grounded Planning2026.032.933.4
GPT 5.2 + SAM3 + ER1VLM Category=Proprietary VLMs, Planning Approach=Decoupled Task Planning + Spatial Grounding2026.031.432.1
Qwen3-VL-4BVLM Category=Open-Source VLMs, Planning Approach=End-to-End Spatially Grounded Planning2026.031.423
Qwen3-VL-4B + ER1VLM Category=Open-Source VLMs, Planning Approach=Decoupled Task Planning + Spatial Grounding2026.030.721
Qwen3-VL-4B + SAM3 + ER1VLM Category=Open-Source VLMs, Planning Approach=Decoupled Task Planning + Spatial Grounding2026.030.720.2
Qwen3-VL-32B + SAM3 + ER1VLM Category=Open-Source VLMs, Planning Approach=Decoupled Task Planning + Spatial Grounding2026.030.733.7
GPT 5.2VLM Category=Proprietary VLMs, Planning Approach=End-to-End Spatially Grounded Planning2026.0302.3
Qwen3-VL-4B + SoMVLM Category=Open-Source VLMs, Planning Approach=Decoupled Task Planning + Spatial Grounding2026.0308.1
InternVL3.5-8BVLM Category=Open-Source VLMs, Planning Approach=End-to-End Spatially Grounded Planning2026.0300