General AI Assistant Tasks on GAIA
93.2AccuracyArgus-35B-A3B (Parallel)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Argus-35B-A3B (Parallel)Category=Parallel Agents, Mode=Parallel2026.05 | 93.2 | — | — | — | |
| Argus-35B-A3B (Solo)Category=Parallel Agents, Mode=Solo2026.05 | 88 | — | — | — | |
| Seed1.82026.02 | 87.4 | — | — | — | |
| Searcher-35B-A3BCategory=Parallel Agents2026.05 | 84.5 | — | — | — | |
| M2CLModel=Qwen-72B, Number of LLMs=642026.02 | 84.2 | — | — | — | |
| MiroThinker-1.7Category=Open-Source Deep Research Agents2026.05 | 82.7 | — | — | — | |
| Gemini-3.1-ProCategory=Proprietary Agents2026.05 | 80.6 | — | — | — | |
| REDSearcherSize=30B-A3B2026.02 | 80.1 | — | — | — | |
| Qwen3.5-35B-A3BCategory=Open-Source Agents2026.05 | 80 | — | — | — | |
| M2CLBackbone=Qwen-72B, Number of LLMs=322026.02 | 79.7 | — | — | — | |
| Seed-2.0-ProCategory=Proprietary Agents2026.05 | 79.6 | — | — | — | |
| Mementocrawl(s)=✓, visual=✓, file=✓, browser=✗, Backbone=O3 + GPT-4.12026.03 | 79.4 | — | — | — | |
| Kimi-K2.6Category=Open-Source Agents2026.05 | 78.6 | — | — | — | |
| GPT-5-Thinking-high2026.02 | 76.7 | — | — | — | |
| Claude-4.6-OpusCategory=Proprietary Agents2026.05 | 75 | — | — | — | |
| Gemini-3-Pro2026.02 | 74.8 | — | — | — | |
| WebSailorV2-30BSize=30B-A3B2026.02 | 74.1 | — | — | — | |
| Tongyi-DeepResearch HeavyCategory=Parallel Agents2026.05 | 72.8 | — | — | — | |
| M2CLModel=Qwen-72B, Number of LLMs=162026.02 | 71.9 | — | — | — | |
| Tongyi DeepResearch-30BSize=30B-A3B2026.02 | 70.9 | — | — | — | |
| Tongyi-DeepResearchCategory=Open-Source Deep Research Agents2026.05 | 70.9 | — | — | — | |
| OpenAI-o32026.02 | 70.5 | — | — | — | |
| GLM-5.0Category=Open-Source Agents2026.05 | 70 | — | — | — | |
| OWLcrawl(s)=✓, visual=✓, file=✓, browser=✓, Backbone=O3-mini + 4o + Claude-S3.72026.03 | 69.7 | — | — | — | |
| WebClipper (Hybrid)Category=Pruning Method, Base Model=Tongyi-DeepResearch, Variant=Hybrid2026.02 | 69.5 | 0.744 | 19.92 | 6,635 | |
| WebClipper (Eff)Category=Pruning Method, Base Model=Tongyi-DeepResearch, Variant=Efficiency2026.02 | 68.4 | 0.76 | 14.44 | 4,756 | |
| Claude-4-SonnetCategory=Close-sourced System2026.02 | 68.3 | — | — | — | |
| M2CLModel=Llama-70B, Number of LLMs=642026.02 | 68.2 | — | — | — | |
| Tongyi-DeepResearchCategory=Open-sourced Agent, Evaluation Environment=Unified2026.02 | 68.2 | 0.733 | 20.56 | 7,378 | |
| OpenAI DeepResearchCategory=Close-sourced System2026.02 | 67.4 | — | — | — | |
| M2CLBackbone=Qwen-72B, Number of LLMs=82026.02 | 67.2 | — | — | — | |
| Prompt ControlCategory=Pruning Method, Base Model=Tongyi-DeepResearch2026.02 | 66.3 | 0.73 | 18.7 | 6,752 | |
| Claude-4.5-sonnet2026.02 | 66 | — | — | — | |
| GLM-4.5 HeavyCategory=Parallel Agents2026.05 | 66 | — | — | — | |
| M2CLBackbone=Llama-70B, Number of LLMs=322026.02 | 65.7 | — | — | — | |
| DeepSeek-V4-Pro-MaxCategory=Open-Source Agents2026.05 | 65.3 | — | — | — | |
| M2CLModel=Llama-70B, Number of LLMs=162026.02 | 61.5 | — | — | — | |
| M2CLBase Model=Qwen-72B, Number of LLMs=42026.02 | 60.3 | — | — | — | |
| MacNetModel=Qwen-72B, Number of LLMs=642026.02 | 59.5 | — | — | — | |
| M2CLBackbone=Llama-70B, Number of LLMs=82026.02 | 58.7 | — | — | — | |
| INFOTREE2026.05 | 56 | — | — | — | |
| MacNetBackbone=Qwen-72B, Number of LLMs=322026.02 | 54.7 | — | — | — | |
| M2CLModel=Qwen-14B, Number of LLMs=642026.02 | 51.8 | — | — | — | |
| Coarse PruneCategory=Pruning Method, Base Model=Tongyi-DeepResearch2026.02 | 51.4 | 0.638 | 15.6 | 4,068 | |
| UIS-Digger (Pangu)crawl(s)=✓, visual=✓, file=✓, browser=✓, Backbone=PanGu-38B2026.03 | 50.5 | — | — | — | |
| MacNetModel=Qwen-72B, Number of LLMs=162026.02 | 50.3 | — | — | — | |
| Tree-GRPO2026.05 | 50.1 | — | — | — | |
| M2CLBase Model=Llama-70B, Number of LLMs=42026.02 | 49.6 | — | — | — | |
| MacNetModel=Llama-70B, Number of LLMs=642026.02 | 49 | — | — | — | |
| M2CLBackbone=Qwen-14B, Number of LLMs=322026.02 | 49 | — | — | — | |
| DebateModel=Qwen-72B, Number of LLMs=642026.02 | 47.6 | — | — | — | |
| UIS-Digger (Qwen)crawl(s)=✓, visual=✓, file=✓, browser=✓, Backbone=Qwen3-32B2026.03 | 47.6 | — | — | — | |
| Kimi-K2-Instruct-0905Category=Open-sourced Agent, Evaluation Environment=Unified2026.02 | 46.9 | 0.625 | 6.45 | 1,281 | |
| DebateBackbone=Qwen-72B, Number of LLMs=322026.02 | 46.6 | — | — | — | |
| MacNetBackbone=Llama-70B, Number of LLMs=322026.02 | 46.5 | — | — | — | |
| MacNetBackbone=Qwen-72B, Number of LLMs=82026.02 | 46.4 | — | — | — | |
| Qwen3-235B-A22B-Instruct-2507Category=Open-sourced Agent, Evaluation Environment=Unified2026.02 | 45.6 | 0.612 | 7.14 | 1,128 | |
| BoNModel=Qwen-72B, Number of LLMs=642026.02 | 44.9 | — | — | — | |
| Flat GRPO2026.05 | 44.8 | — | — | — | |
| DebateModel=Qwen-72B, Number of LLMs=162026.02 | 44.1 | — | — | — | |
| M2CLModel=Qwen-14B, Number of LLMs=162026.02 | 43.9 | — | — | — | |
| BoNBackbone=Qwen-72B, Number of LLMs=322026.02 | 43.9 | — | — | — | |
| MacNetModel=Llama-70B, Number of LLMs=162026.02 | 43.8 | — | — | — | |
| DyLANModel=Qwen-72B, Number of LLMs=642026.02 | 43.7 | — | — | — | |
| BoNModel=Qwen-72B, Number of LLMs=162026.02 | 43.3 | — | — | — | |
| GPTSwarmModel=Qwen-72B, Number of LLMs=642026.02 | 43 | — | — | — | |
| DebateBackbone=Qwen-72B, Number of LLMs=82026.02 | 42.9 | — | — | — | |
| DyLANBackbone=Qwen-72B, Number of LLMs=322026.02 | 42.9 | — | — | — | |
| MacNetModel=Qwen-14B, Number of LLMs=642026.02 | 42.7 | — | — | — | |
| GPTSwarmBackbone=Qwen-72B, Number of LLMs=322026.02 | 42.7 | — | — | — | |
| M2CLBackbone=Qwen-14B, Number of LLMs=82026.02 | 42 | — | — | — | |
| M2CLModel=Qwen-7B, Number of LLMs=642026.02 | 42 | — | — | — | |
| MacNetBase Model=Qwen-72B, Number of LLMs=42026.02 | 42 | — | — | — | |
| GPTSwarmModel=Qwen-72B, Number of LLMs=162026.02 | 41.3 | — | — | — | |
| BoNBackbone=Qwen-72B, Number of LLMs=82026.02 | 41.2 | — | — | — | |
| DyLANModel=Qwen-72B, Number of LLMs=162026.02 | 41.1 | — | — | — | |
| MacNetBackbone=Llama-70B, Number of LLMs=82026.02 | 40.7 | — | — | — | |
| DebateBase Model=Qwen-72B, Number of LLMs=42026.02 | 40.7 | — | — | — | |
| DyLANBackbone=Qwen-72B, Number of LLMs=82026.02 | 40.4 | — | — | — | |
| GPTSwarmBackbone=Qwen-72B, Number of LLMs=82026.02 | 40.3 | — | — | — | |
| M2CLModel=Llama-14B, Number of LLMs=642026.02 | 40.3 | — | — | — | |
| MacNetBackbone=Qwen-14B, Number of LLMs=322026.02 | 39.8 | — | — | — | |
| M2CLBackbone=Qwen-7B, Number of LLMs=322026.02 | 39.7 | — | — | — | |
| DeepSeek-R1-671BCategory=Open-sourced Agent, Evaluation Environment=Unified2026.02 | 39.2 | 0.557 | 4.01 | 1,468 | |
| GPTSwarmBase Model=Qwen-72B, Number of LLMs=42026.02 | 39 | — | — | — | |
| BoNBase Model=Qwen-72B, Number of LLMs=42026.02 | 38.6 | — | — | — | |
| M2CLBackbone=Llama-14B, Number of LLMs=322026.02 | 38.5 | — | — | — | |
| DyLANBase Model=Qwen-72B, Number of LLMs=42026.02 | 38.2 | — | — | — | |
| M2CLBase Model=Qwen-14B, Number of LLMs=42026.02 | 37.2 | — | — | — | |
| WebExplorerCategory=Open-sourced Agent2026.02 | 37.2 | 0.521 | 12.88 | 3,560 | |
| M2CLModel=Qwen-7B, Number of LLMs=162026.02 | 36.5 | — | — | — | |
| MacNetBase Model=Llama-70B, Number of LLMs=42026.02 | 36.5 | — | — | — | |
| M2CLModel=Llama-14B, Number of LLMs=162026.02 | 36.2 | — | — | — | |
| MacNetModel=Qwen-14B, Number of LLMs=162026.02 | 35.5 | — | — | — | |
| BoNModel=Llama-70B, Number of LLMs=642026.02 | 33.9 | — | — | — | |
| M2CLBackbone=Llama-14B, Number of LLMs=82026.02 | 33.6 | — | — | — | |
| M2CLBackbone=Qwen-7B, Number of LLMs=82026.02 | 33.6 | — | — | — | |
| BoNModel=Llama-70B, Number of LLMs=162026.02 | 33.2 | — | — | — | |
| DebateModel=Llama-70B, Number of LLMs=162026.02 | 33 | — | — | — | |
| BoNBackbone=Llama-70B, Number of LLMs=82026.02 | 32.9 | — | — | — |