# SearchSwarm: Towards Delegation Intelligence in Agentic LLMs for Long-Horizon Deep Research > Training a 30B-A3B model on harness-elicited delegation trajectories yields state-of-the-art on long-horizon benchmarks, rivaling 10x larger models. - **Source:** [arXiv](https://arxiv.org/abs/2606.09730) - **Published:** 2026-06-11 - **Permalink:** https://picx.dev/p/aFopgw - **Whiteboard:** https://picx.dev/p/aFopgw/image ## Summary ## Summary (Overview) - **Problem**: Long-horizon agent tasks generate unbounded context demands, but LLM context windows are inherently finite. Existing passive context management (truncation, summarization) lacks prior planning. - **Proposal**: Introduce *delegation intelligence* — the main agent decomposes tasks, dispatches bounded subtasks to subagents via `call_sub_agent`, and integrates condensed reports. A harness guides optimal delegation behavior at inference. - **Data Synthesis**: Harness-guided trajectories that encode correct delegation decisions are used as supervised fine-tuning (SFT) data to internalize delegation behavior into model weights. - **Results**: SearchSwarm-30B-A3B achieves 68.1 on BrowseComp, 73.3 on BrowseComp-ZH, 82.5 on GAIA, and 80.8 on xbench-DeepSearch — best among all models of comparable scale and competitive with models over 10× larger. - **Open Source**: Harness, model weights, and training data are released to facilitate future research. ## Introduction and Theoretical Foundation Large language models (LLMs) increasingly act as agents for complex, multi-step tasks whose information needs grow without bound. Yet context windows remain finite, creating a fundamental tension. Early context management strategies — summarization after length thresholds, discarding old tool outputs — are *passive*: they react after budget exhaustion without prior planning. An alternative paradigm is **delegation**: a main agent decomposes the task in advance, dispatches bounded subtasks to subagents, and receives only condensed results, actively preserving context budget. The required capability is termed **delegation intelligence**: the ability to decompose, determine when and what to delegate, and integrate results. Naturally occurring text rarely exhibits explicit multi-agent coordination, making training data for delegation intelligence scarce. The paper presents a structured recipe to synthesize such data in the **deep research** domain, a representative long-horizon agent task. The core idea: design a *harness* that elicits high-quality delegation at inference time, then use the resulting trajectories as SFT data to internalize the behavior. ## Methodology ### Formulation The deep research task is modelled as a multi-turn ReAct (Yao et al., 2022) interaction. At step \( t \): - **Thought** \( \tau_t \): internal reasoning - **Action** \( a_t \): tool call (including `call_sub_agent`) - **Observation** \( o_t \): environment result Trajectory: $$ H_T = \langle q, (\tau_0, a_0, o_0), \ldots, (\tau_T, a_T, o_T), y \rangle \tag{1} $$ Policy: $$ \tau_t, a_t \sim \pi(\cdot \mid q, H_{t-1}) \tag{2} $$ When \( a_t = \text{call\_sub\_agent}(b) \), the brief \( b \) triggers an independent sub‑trajectory: $$ H_{\text{sub}} = \langle b, (\tau^s_0, a^s_0, o^s_0), \ldots, (\tau^s_S, a^s_S, o^s_S), r \rangle \tag{3} $$ and the main agent receives \( o_t = r \) (the condensed report). Intermediate steps are not visible. ### Harness Design The harness consists of a tool set and system prompts that guide an LLM toward high-quality delegation via four principles: | Principle | Description | |-----------|-------------| | **Encouraging delegation** | Main agent should offload token‑expensive but cognitively shallow information gathering to subagents, reserving its own context for high‑level coordination and verification. | | **Comprehensive briefing** | Each brief must include subtask description, why it matters, what is already known, what remains uncertain, and directions already ruled out — so subagents operate with full context. | | **Main agent retains core judgment** | Only the main agent has a complete view across subtasks; it must independently decide hypotheses, termination, and adjudication. Subagents focus on evidence gathering. | | **Citation‑grounded reporting** | Subagent reports attach inline citations to every important conclusion. The main agent propagates these to its final answer, ensuring end‑to‑end traceability. | Tools include `search`, `visit`, `google_scholar`, `python`, and `call_sub_agent`. Single‑level delegation (subagents cannot delegate further). ### Supervised Fine‑Tuning **Data collection**: Queries from RedSearcher and OpenSeeker datasets. The model executes deep research under harness guidance. Two configurations: 1. Same model as main and subagent – both trajectories retained. 2. Stronger main agent + weaker subagent – only main agent trajectories retained (more deliberate decomposition and verification). **Filtering**: Retain only main trajectories with correct final answers. Remove repeated tool calls, hallucinated citations, tool misuse. **Training objective**: next‑token prediction with environment masking: $$ \mathcal{L} = -\sum_{t=1}^{T} \sum_{j=1}^{|a_t|} \log p_\theta \left( a_t^{(j)} \mid a_t^{(