| MiroThinker-1.7 & H1: Towards Heavy-Duty Research Agents via Verification | | 84.4 | MiroThinker-H1 | 2026-03-16 |
| dots3-note Preview: A Small but Mighty Step Toward Long-Horizon Agency in Real Life | ✓ Link | 75.78 | dots3-note Preview | 2026-08-13 |
| Masking Stale Observations Helps Search Agents -- Until It Doesn't: A Regime Map and Its Mechanism | ✓ Link | 73.4 | DS-V4-Flash-Max + Serper API (stale-observation masking, K=5) | 2026-05-29 |
| AgentSwing: Adaptive Parallel Context Management Routing for Long-Horizon Web Agents | | 71.3 | AgentSwing (DeepSeek-v3.2) | 2026-03-29 |
| DeepSeek-V3.2 Technical Report | ✓ Link | 65.0 | DeepSeek-V3.2 | 2025-12-02 |
| ERNIE 5.0 Technical Report | | 64.71 | ERNIE 5.0 | 2026-02-04 |
| SAM: State-Adaptive Memory for Long-Horizon Reasoning Agent | ✓ Link | 64.2 | GLM-4.7 with SAM | 2026-05-23 |
| Table-as-Search: Formulate Long-Horizon Agentic Information Seeking as Table Completion | ✓ Link | 63.7 | GPT-5 Medium-Think (TaS) | 2026-02-06 |
| OpenSeeker-v2: Pushing the Limits of Search Agents with Informative and High-Difficulty Trajectories | ✓ Link | 58.1 | OpenSeeker-v2 (30B, SFT) | 2026-05-05 |
| IterSynth: Rethinking Deep Search Agents via Role-Decoupled Iterative Synthesis | ✓ Link | 55.4 | IterSynth-8B | 2026-09-24 |
| ABSeeker: Training Long-Horizon Search Agents via Answer-Backtracked Credit Assignment | | 52.9 | ABSeeker (Qwen3.5-4B, context management) | 2026-08-05 |
| OpenSeeker: Democratizing Frontier Search Agents by Fully Open-Sourcing Training Data | ✓ Link | 48.4 | OpenSeeker-v1-30B-SFT | 2026-03-16 |
| Marco DeepResearch: Unlocking Efficient Deep Research Agents via Verification-Centric Design | ✓ Link | 47.1 | Marco-DeepResearch-8B | 2026-03-30 |
| Mind DeepResearch Technical Report | | 45.7 | MindDR-v1.5-30B-A3B | 2026-04-17 |
| MiniCPM5-2B | ✓ Link | 43.5 | MiniCPM5-2B | 2026-09-07 |
| Nested Browser-Use Learning for Agentic Information Seeking | | 42.6 | NestBrowse-30B-A3B | 2025-12-29 |
| LightThinker++: From Reasoning Compression to Memory Management | | 36.9 | LThinker++ (Qwen3-30B-A3B) | 2026-04-04 |
| Learning Agentic Policy from Action Guidance | ✓ Link | 26.64 | ActGuide-RL (Qwen3-8B) | 2026-05-12 |
| Let It Flow: Agentic Crafting on Rock and Roll, Building the ROME Model within an Open Agentic Learning Ecosystem | ✓ Link | 14.19 | ROME (30B MoE, activated 3B) | 2025-12-31 |