Tag
SWE-Shift 是一个评估编码智能体在执行过程中遭遇环境、依赖与需求变化时表现的新基准，包含120个场景和12类变化。在评测的17个编码智能体系统中，Shifted 条件下的成功率均低于 Clean 条件，GPT-6 Astra 从 67.35% 骤降至 24.79%。
DynaWeave 是一个自动化合成框架，可将静态任务转化为含外生事件的可验证动态场景，用于训练 LLM 智能体适应环境变化；使用其生成的 2,095 条轨迹对 Qwen3.5 进行监督微调后，在 CostBench 上任务完成率提升 67.34 个百分点，并在 ToolMaze 上超越 Gemini-3.1-Pro-Preview。
EnvHarness is a programmable wrapper framework that dynamically adapts static environments to enhance LLM agent training, outperforming existing methods by providing superior optimization signals for reinforcement learning across multiple benchmarks.
dots3-note Preview is an AI model targeting autonomous agency and dynamic real-world environments with a 512K context window.
PACE-Bench introduces a simulator-grounded benchmark for evaluating self-evolving agents on physics adaptation tasks involving iterative code redesign after environmental mutations, revealing that mechanism redesign is a major bottleneck compared to parameter inference.
Seed IQ demonstrates advanced real-time perception, reasoning, and adaptation in dynamic environments by navigating Doom II, potentially surpassing static benchmarks like ARC AGI 3.
EvoArena introduces a benchmark for evaluating LLM agents in dynamic environments with progressive updates across terminal, software, and social domains, while EvoMem proposes a patch-based memory paradigm that records structured evolution; experiments show current agents achieve only 39.6% accuracy on EvoArena, and EvoMem yields average gains of 1.5% on the benchmark and improvements on GAIA and LoCoMo.
SkillHarness is a framework that enables computer-use agents to safely learn and execute skills in dynamic environments by incorporating safety constraints and adaptive skill selection mechanisms, reducing unsafe rates by 57.1%.