List your company

Web Navigation Task Success on MIND2WEB ONLINE (test)

67Task Success Rate (Overall)Orchard-GUI-4B (SFT + RL)
7.222.72538.2553.775Feb 2, 2026Feb 18, 2026Mar 7, 2026Mar 24, 2026Apr 10, 2026Apr 27, 2026May 14, 2026

Evaluation results

MethodLinks
Orchard-GUI-4B (SFT + RL)# Steps=30, # Tasks=2.6k2026.0567———
NavigatorMain Model=n1-preview-11-2025, Provider=Yutori, Open Source=false2026.0264.78462.248.7
Qwen3-VL-235B-A22B-Thinking# Steps=302026.0563.7———
OperatorMain Model=OpenAI Computer-Using Agent, Provider=OpenAI, Open Source=false2026.0258.373.559.439.2
OpenAI computer-use-preview# Steps=1002026.0558.3———
GPT-5 (SoM)# Steps=1002026.0557.7———
Google Computer UseMain Model=Gemini 2.5 Computer Use, Provider=Google DeepMind, Open Source=false2026.0257.377.155.245.9
Gemini computer-use-preview# Steps=1002026.0557.3———
o3 (SoM)# Steps=1002026.0555.4———
AVENIR-WEBMain Model=Gemini 3 Pro, Provider=Ours, Open Source=true2026.0253.774.154.630.3
ACT-1-20250814Main Model=o3-2025-04-16 and Claude-sonnet-4-20250514, Provider=Enhans, Open Source=false2026.0252.771.152.432.4
WebNavigatorModel=Gemini-2.5-Pro, Act #=62026.0352.7———
Claude Computer Use 3.7Main Model=Claude-3.7-sonnet-20250219, Provider=Anthropic, Open Source=false2026.0247.375.941.327
Orchard-GUI-4B-SFT# Steps=30, # Tasks=0.4k2026.0547———
Gemini-3-flash (Axtree)# Steps=1002026.0544.8———
GPT-5 (Axtree)# Steps=302026.0541.9———
WebNavigatorModel=GPT-4o, Act #=62026.0341.3———
WebNavigatorModel=Qwen3-VL-32B-Instruct, Act #=62026.0339.7———
ACT-1-20250703Main Model=o3-2025-04-16 and Claude-sonnet-4-20250514, Provider=Enhans, Open Source=false2026.0239.553.739.224.3
WebNavigatorModel=Claude-Sonnet-4, Act #=62026.0338.7———
MolmoWeb-8B# Steps=100, # Tasks=>278.5K2026.0535.3———
WebDreamerModel=GPT-4o, Act #=122026.0335———
Gemini-3-flash (Axtree)# Steps=302026.0534.8———
GPT-4o (SoM)# Steps=1002026.0534.6———
Fara-7B# Steps=100, # Tasks=>123.2k2026.0534.1———
GLM-4.1V-9B-Thinking# Steps=1002026.0533.9———
Qwen3-VL-4B-Thinking# Steps=302026.0532———
UI-TARS-1.5-7B# Steps=1002026.0531.3———
MolmoWeb-4B# Steps=100, # Tasks=>278.5k2026.0531.3———
SeeActMain Model=gpt-4o-2024-08-06, Provider=OSU, Open Source=true2026.023051.8289.5
SeeActBackbone=GPT-42026.0428———
Agent-EMain Model=gpt-4o-2024-08-06, Provider=Emergence AI, Open Source=true2026.022751.823.16.8
Agent-eModel=GPT-4o, Act #=82026.0327———
PageGuideBackbone=google/gemini-3-flash-preview2026.0426.57———
Browser UseMain Model=gpt-4o-2024-08-06, Provider=Browser Use, Open Source=true2026.022644.623.110.8
AVENIR-WEBMain Model=Qwen-3-VL-8B, Provider=Ours, Open Source=true2026.0225.74223.811.8
PageGuideBackbone=google/gemini-2.5-flash-preview2026.0424.48———
Claude 3.5Main Model=Claude-3-5-sonnet-20241022, Provider=Anthropic, Open Source=false2026.022451.816.18.1
PageGuideBackbone=google/gemini-3-flash-preview2026.0418.18———
PageGuideBackbone=google/gemini-2.5-flash-preview2026.0414.29———
SeeActBackbone=GPT-42026.049.5———
InfiniteWebTraining Tasks=2002026.01—54.318.27.9
InfiniteWebTraining Tasks=4002026.01—48.227.35.3
InfiniteWebTraining Tasks=6002026.01—56.823.19.2
Original BaselineTraining Tasks=Orig.2026.01—46.918.95.3