List your company

Omni-modal agentic tasks on OmniGAIA

75Pass@1 (Overall)GPT-5.4 xHigh
-1.54418.32838.258.072Feb 26, 2026Mar 15, 2026Apr 1, 2026Apr 19, 2026May 6, 2026May 23, 2026Jun 10, 2026

Evaluation results

MethodLinks
GPT-5.4 xHighHarness=Codex, Reasoning effort=xHigh2026.0575—————————827564.1
Orchestra-o1-GPT-5Model Accessibility=Proprietary, Backbone=GPT-5, Agent Framework=Orchestra-o12026.0672.872.569.475.86483.863.969.773.183.3———
Claude Opus 4.6Harness=CC2026.0568.6—————————74.669.457.7
OmniAgentBackbone=GPT-5.4 High2026.0566.4—————————77.963.155.1
Gemini 3.1 Pro2026.0566.1—————————77.863.852.6
GPT-5.4 High§Harness=Codex, Direct image read=disabled2026.0564.4—————————70.563.856.4
Claude Sonnet 4.6Harness=CC2026.0563.3—————————73.862.548.8
Gemini-3-Pro# Params=-2026.0262.565.259.262.17278.452.848.542.388.978.761.938.5
Gemini-3-ProModel Accessibility=Proprietary, Agent Framework=ReAct2026.0662.565.259.262.17278.452.848.542.388.9———
GPT-5.4 HighHarness=Codex, Reasoning effort=High2026.0561.4—————————70.46050
GPT-5.4 MediumHarness=Codex, Reasoning effort=Medium2026.0556.4—————————66.45543.6
Kimi K-2.5Harness=CC2026.0555.6—————————72.150.639.7
GPT-5.4 LowHarness=Codex, Reasoning effort=Low2026.0555—————————63.153.844.9
Codex-GPT-5.4 LowSetting=Strong Baseline2026.0555—————————63.153.844.9
Gemini-3-Flash# Params=-2026.0251.750.757.144.84859.555.654.638.561.167.246.937.2
Gemini 3 Flash2026.0551.7—————————67.246.937.2
Gemini-3-FlashModel Accessibility=Proprietary, Agent Framework=ReAct2026.0651.750.757.144.84859.555.654.638.561.1———
Code-X (SFT)Model=Qwen3.6-27B, Thinking budget=4K2026.0543.3—————————504528.2
Agent-OmniBackbone=GPT-5.4 High2026.0541.1—————————54.141.919.2
AOrchestra-GPT-5Model Accessibility=Proprietary, Backbone=GPT-5, Agent Framework=AOrchestra2026.064034.840.856.13251.42542.430.822.2———
Qwen 3.5 Omni Plus2026.0538.7—————————43.539.429
Direct MLLMModel=Qwen3.6-27B, Thinking budget=4K2026.0538.2—————————44.338.128.6
MiniMax-M2.7‡Harness=CC, Note=Uses own MCP tool for image reading2026.0533.3—————————38.534.423.1
Gemini-2.5-Pro# Params=-2026.0230.823.228.632.82032.441.742.426.933.341.826.921.8
Gemini-2.5-ProModel Accessibility=Proprietary, Agent Framework=ReAct2026.0630.823.228.632.82032.441.742.426.933.3———
Orchestra-o1-8BModel Accessibility=Open-Source, Backbone=Qwen3-8B, Agent Framework=Orchestra-o12026.063021.732.737.91229.716.745.538.538.9———
Code-X (RL)Model=Qwen3.5-9B, Thinking budget=4K2026.0528.1—————————32.826.324.4
Mimo V2 Omni2026.0525.8—————————33.922.519.1
Baseline (direct image)Model=Qwen3.6-27B, Thinking budget=4K2026.0524.4—————————32.818.823.1
Code-X (SFT)Model=Qwen3.5-9B, Thinking budget=4K2026.0523.1—————————302117
OmniAtlas-Qwen-3# Params=30B-A3B2026.0220.810.130.629.93218.916.712.111.527.831.118.89
OmniAtlas-Omni-30B-A3BSetting=Strong Baseline2026.0520.8—————————31.118.89
OmniAtlas-Qwen3-30B-A3BModel Accessibility=Open-Source, Agent Framework=ReAct2026.0620.810.130.629.93218.916.712.111.527.8———
Direct MLLMModel=Qwen3.5-9B, Thinking budget=4K2026.0520—————————2316.223.1
Baseline (direct image)Model=Qwen3.5-9B, Thinking budget=4K2026.0518.3—————————23.814.417.9
Qwen-3-Omni# Params=30B-A3B2026.0213.38.714.311.92810.813.99.115.422.219.710.69
OmniAtlas-Qwen-2.5# Params=7B2026.0213.38.718.416.4416.222.237.722.222.111.33.9
Qwen 3 Omni2026.0513.3—————————19.710.69
Qwen3-Omni-30B-A3BModel Accessibility=Open-Source, Agent Framework=ReAct2026.0613.38.714.311.92810.813.99.115.422.2———
OmniAtlas-Qwen2.5-7BModel Accessibility=Open-Source, Agent Framework=ReAct2026.0613.38.718.416.4416.222.237.722.2———
LongCat-Flash-Omni# Params=560B-A27B2026.0211.18.710.216.41210.88.36.111.516.716.49.46.4
LongCat-Flash-Omni-560B-A27BModel Accessibility=Open-Source, Agent Framework=ReAct2026.0611.18.710.216.41210.88.36.111.516.7———
OmniAtlas-Qwen-2.5# Params=3B2026.0210.34.412.216.7416.211.1311.511.113.9105.1
OmniAtlas-Qwen2.5-3BModel Accessibility=Open-Source, Agent Framework=ReAct2026.0610.34.412.216.7416.211.1311.511.1———
Gemini-2.5-Flash-Lite# Params=-2026.028.65.88.214.9410.88.36.13.911.19.88.17.7
Gemini-2.5-Flash-LiteModel Accessibility=Proprietary, Agent Framework=ReAct2026.068.65.88.214.9410.88.36.13.911.1———
Ming-Flash-Omni# Params=100B-A6B2026.028.35.88.210.4128.15.66.111.511.112.37.53.8
Ming-Flash-Omni-100B-A6BModel Accessibility=Open-Source, Agent Framework=ReAct2026.068.35.88.210.4128.15.66.111.511.1———
Ming-Lite-Omni-1.5# Params=20B-A3B2026.023.92.96.11.545.42.86.17.75.64.93.82.6
Ming-Lite-Omni-1.5-20B-A3BModel Accessibility=Open-Source, Agent Framework=ReAct2026.063.92.96.11.545.42.86.17.75.6———
Qwen-2.5-Omni# Params=7B2026.023.61.54.17.5402.807.75.68.21.31.3
Qwen2.5-Omni-7BModel Accessibility=Open-Source, Agent Framework=ReAct2026.063.61.54.17.5402.807.75.6———
MiniCPM-O-2.6# Params=8B2026.023.12.921.502.78.333.85.63.32.53.8
MiniCPM-O-2.6-8BModel Accessibility=Open-Source, Agent Framework=ReAct2026.063.12.921.502.78.333.85.6———
Baichuan-Omni-1.5# Params=8B2026.022.82.94.1342.7033.804.92.50
Baichuan-Omni-1.5-8BModel Accessibility=Open-Source, Agent Framework=ReAct2026.062.82.94.1342.7033.80———
Qwen-2.5-Omni# Params=3B2026.021.4024.500003.901.61.90
Qwen2.5-Omni-3BModel Accessibility=Open-Source, Agent Framework=ReAct2026.061.4024.500003.90———