Guannan Zhang

Occamy-1.0, a cost-efficient co-work model obtained by further training the post-trained Qwen3.6-35B-A3B checkpoint, is presented and placed at the low-cost knee of the observed cost--performance Pareto frontier.

AdaVDR achieves superior performance in video deep research by dynamically adapting its tool use based on the specific capabilities of the model and the nature of the task.

LLM agents exhibit a staggering ninefold difference in performance when navigating the complexities of a simulated marketplace, yet still fall short of human-designed strategies.

NOTES reduces design dimensionality from 256 to 25 while achieving over 95% efficiency in inverse design tasks, outpacing conventional methods.

Agentic models can learn to trust their "gut" and rely less on external tools, leading to faster and more accurate reasoning.