List your company

Web Navigation on Mind2Web Cross-Domain (EA, AF1, StepSR, TaskSR)

68.4Element Accuracy (EA)GUICrafter-7B
2.98419.96736.9553.933Sep 29, 2025Nov 13, 2025Dec 29, 2025Feb 12, 2026Mar 30, 2026May 14, 2026Jun 29, 2026

Evaluation results

MethodLinks
GUICrafter-7BModel Size=About 7B2026.0668.491.362.2—
GUICrafter-7B (only Stage2)Model Size=About 7B, Training Stage=Stage 22026.0666.789.760.1—
UI-TARS-7BModel Size=About 7B2026.0666.690.960.5—
GUI-R1-7B (v2)Model Size=About 7B2026.0665.988.659—
FluxMemBase Model=GPT-4.1-mini, Setting=Filtered (manual removal of non-golden HTML elements)2026.0565.275.355.37.2
FluxMemBase Model=Gemini-2.5-flash, Setting=Filtered (manual removal of non-golden HTML elements)2026.0564.870.354.67
GPT-4o + Aria-UIAgent Framework=Aria-UI [47]2026.0661.4———
GUICrafter-3BModel Size=About 3B2026.0660.790.652.9—
Aguvis-7BModel Size=About 7B2026.0660.489.256.6—
Qwen2.5-VL-7B (Fine-tuned)Model Size=About 7B, Protocol=Fine-tuned2026.0660.289.256.4—
UI-TARS-2BModel Size=About 3B2026.0658.889.652.3—
FluxMemBase Model=Gemini-2.5-flash, Setting=Realistic (without any pre-filtering)2026.0558.570.155.95
FluxMemBase Model=GPT-4.1-mini, Setting=Realistic (without any pre-filtering)2026.0556.269.452.14.3
GUICrafter-3B (only Stage2)Model Size=About 3B, Training Stage=Stage 22026.0656.286.750.7—
GUI-R1-3B (v2)Model Size=About 3B2026.0654.785.849.8—
Upper BoundBackbone=Qwen2.5VL-3B2026.0353.2560.2748.89—
DivPruneBackbone=Qwen2.5VL-3B2026.0352.8958.9648.54—
PDropBackbone=Qwen2.5VL-3B2026.0352.8159.9848.42—
RandomBackbone=Qwen2.5VL-3B2026.0352.6558.7848.13—
FASTVBackbone=Qwen2.5VL-3B2026.0352.5558.348.28—
SparseVLMBackbone=Qwen2.5VL-3B2026.0352.4458.4748.16—
AWMBase Model=Gemini-2.5-flash, Setting=Filtered (manual removal of non-golden HTML elements)2026.05526747.93.4
DartBackbone=Qwen2.5VL-3B2026.0351.9759.3347.75—
GUICrafter-7B (only Stage1)Model Size=About 7B, Training Stage=Stage 12026.065158.846.1—
AWMBase Model=GPT-4.1-mini, Setting=Filtered (manual removal of non-golden HTML elements)2026.0550.569.343.92
Upper BoundBackbone=Qwen2VL-2B2026.0347.657.8643.99—
FASTVBackbone=Qwen2VL-2B2026.0347.3457.2143.71—
PDropBackbone=Qwen2VL-2B2026.0347.258.0743.66—
Qwen2.5-VL-3B (Fine-tuned)Model Size=About 3B, Protocol=Fine-tuned2026.064755.743.2—
RandomBackbone=Qwen2VL-2B2026.0346.757.0943.32—
DivPruneBackbone=Qwen2VL-2B2026.0346.6657.8843.26—
DartBackbone=Qwen2VL-2B2026.0346.6257.2543.04—
GPT-4o + UGroundAgent Framework=UGround [10]2026.0646.6———
GPT-4V + OmniParserAgent Framework=OmniParser [22]2026.0645.585.742—
SparseVLMBackbone=Qwen2VL-2B2026.0345.285741.68—
GUI-R1-7B (v1)Model Size=About 7B2026.0644.64735.7—
ReasoningBankBackbone=Gemini-2.5-pro2025.0942.845.238.11.7
UI-R1-3BModel Size=About 3B2026.0642.5———
HMTBackbone=GPT-42026.0342.348.837.11.9
AWMBase Model=Gemini-2.5-flash, Setting=Realistic (without any pre-filtering)2026.0541.660.238.11.5
AWMSetting=online2026.0340.946.335.51.7
ReasoningBankBackbone=Gemini-2.5-flash2025.0940.641.336.61.6
ReasoningBankBase Model=Gemini-2.5-flash, Setting=Filtered (manual removal of non-golden HTML elements)2026.0540.641.336.61.6
GUICrafter-3B (only Stage1)Model Size=About 3B, Training Stage=Stage 12026.0640.437.536.7—
ShowUI-2B (Fine-tuned)Model Size=About 3B, Protocol=Fine-tuned2026.0639.486.835.2—
GUI-R1-3B (v1)Model Size=About 3B2026.0638.935.932—
SynapseBackbone=Gemini-2.5-pro2025.0938.539.835.61.5
No MemoryBackbone=Gemini-2.5-pro2025.0937.937.7351.4
AWMBackbone=Gemini-2.5-pro2025.0937.338.134.41.2
AWMSetting=offline2026.0336.441.632.60.7
Qwen2.5-VL-7B (Zero-Shot)Model Size=About 7B, Protocol=Zero-Shot2026.0636.440.425.3—
SynapseBackbone=Gemini-2.5-flash2025.0936.338.532.41.1
No MemoryBackbone=Gemini-2.5-flash2025.0935.837.931.91
GPT-4 (Text-only)Input Modality=Text-only2026.0635.461.929.7—
GPT-4o + SeeClickAgent Framework=SeeClick [6]2026.0633.5———
AWMBackbone=Gemini-2.5-flash2025.0933.336.530.10.7
AWMBase Model=GPT-4.1-mini, Setting=Realistic (without any pre-filtering)2026.0530.356.924.91
No MemoryBase Model=GPT-4.1-mini, Setting=Realistic (without any pre-filtering)2026.0530.155.924.60.7
Qwen2.5-VL-3B (Zero-Shot)Model Size=About 3B, Protocol=Zero-Shot2026.0628.834.122.6—
GPT-4 (SOM)Input Modality=SOM2026.0627—23.7—
ShowUI-2B (Zero-Shot)Model Size=About 3B, Protocol=Zero-Shot2026.0624.483.921.4—
No MemoryBase Model=Gemini-2.5-flash, Setting=Realistic (without any pre-filtering)2026.0522.137.218.61.8
MindAct2026.0321.652.826.42
GPT-4oCategory=Proprietary Model2026.065.586.44.5—