Scaling Agents via Continual Pre-training
This addresses the problem of improving autonomous agent capabilities for complex problem-solving in AI, though it appears incremental as it builds upon existing pre-training methods.
The paper tackles the underperformance of post-training approaches for agentic tasks by identifying a root cause in optimization tensions and proposes Agentic Continual Pre-training to build agentic foundation models, resulting in state-of-the-art performance on benchmarks such as 39.9% on BrowseComp-en and 31.5% Pass@1 on HLE.
Large language models (LLMs) have evolved into agentic systems capable of autonomous tool use and multi-step reasoning for complex problem-solving. However, post-training approaches building upon general-purpose foundation models consistently underperform in agentic tasks, particularly in open-source implementations. We identify the root cause: the absence of robust agentic foundation models forces models during post-training to simultaneously learn diverse agentic behaviors while aligning them to expert demonstrations, thereby creating fundamental optimization tensions. To this end, we are the first to propose incorporating Agentic Continual Pre-training (Agentic CPT) into the deep research agents training pipeline to build powerful agentic foundational models. Based on this approach, we develop a deep research agent model named AgentFounder. We evaluate our AgentFounder-30B on 10 benchmarks and achieve state-of-the-art performance while retains strong tool-use ability, notably 39.9% on BrowseComp-en, 43.3% on BrowseComp-zh, and 31.5% Pass@1 on HLE.