| Qwen3.8-Max: A New Bar for Coding and Cowork | ✓ Link | 86.1 | Qwen3.8-Max | 2026-08-03 |
| Kimi K3: Open Frontier Intelligence | ✓ Link | 84.8 | Kimi K3 Max | 2026-07-27 |
| Qwen3.8-27B | ✓ Link | 84.3 | Qwen3.8-27B | 2026-08-05 |
| NexForge: Scaling Agent Capabilities through Requirement-Driven Task Synthesis for LLMs | ✓ Link | 82.2 | Nex-N2.5-Pro | 2026-07-17 |
| MiMo-V2.6: Scaling Reinforcement Learning Towards Self-Improvement | ✓ Link | 82.0 | MiMo-V2.6-Pro-RL | 2026-10-08 |
| Qwen-UI-Agent Technical Report: Toward Next-Generation Real-World Centric Foundation GUI Agents | | 79.5 | Qwen-UI-Agent (27B) | 2026-07-30 |
| VLAA-GUI: Knowing When to Stop, Recover, and Search, A Modular Framework for GUI Automation | ✓ Link | 77.45 | VLAA-GUI (Opus 4.6) | 2026-04-23 |
| UI-Mate: Advancing Open-Weight Foundation GUI Agents with In-Context Demonstrations | ✓ Link | 77.0 | UI-Mate-27B | 2026-08-16 |
| CUA-Gym: Scaling Verifiable Training Environments and Tasks for Computer-Use Agents | ✓ Link | 72.6 | CUA-GYM-A17B | 2026-05-25 |
| GTA1: GUI Test-time Scaling Agent | ✓ Link | 63.4 | GTA1-32B + GPT-5 planner/judge | 2025-07-08 |
| Introducing Claude Sonnet 4.5 | | 61.4 | Claude Sonnet 4.5 | 2025-09-29 |
| Priming Base Models for Agentic Post-Training: An Open Midtraining Recipe | | 51.14 | NVIDIA Nemotron 3.5 Super VL 120B-A12B BF16 | 2026-10-15 |
| OpenCUA: Open Foundations for Computer-Use Agents | ✓ Link | 45.0 | OpenCUA-72B | 2025-08-12 |
| Anticipatory Planning for Multimodal AI Agents | | 41.2 | TraceR1 (Qwen3-VL-32B-Thinking) | 2026-03-17 |
| OpenCUA: Open Foundations for Computer-Use Agents | ✓ Link | 34.8 | OpenCUA-32B | 2025-08-12 |
| LLaDA-UI: Bringing Block-wise Diffusion to Vision-Language GUI Agents | ✓ Link | 29.4 | LLaDA-UI | 2026-09-09 |
| OpenCUA: Open Foundations for Computer-Use Agents | ✓ Link | 23.06 | OpenCUA-Qwen2-7B | 2025-08-12 |