SearchSwarm: New AI Model Improves Task Delegation for Long-Horizon Research
Researchers have introduced SearchSwarm, a framework that trains large language models to decompose and delegate long-horizon research tasks to specialized subagents, achieving state-of-the-art results on deep research benchmarks. The work addresses a fundamental limitation of LLMs — finite context windows — by having a main agent distribute subtasks and integrate only summarized results, preserving its context budget. The approach is notable for synthesizing training data for 'delegation intelligence,' a capability largely absent from existing open-source research.
Researchers have released SearchSwarm, a system designed to give large language models what they term 'delegation intelligence' — the ability to break down complex, open-ended tasks, assign subtasks to subagents, and synthesize their summarized outputs into a coherent workflow. The core challenge motivating the work is that real-world long-horizon tasks can require more information than any single model's context window can hold, making multi-agent decomposition a practical necessity. The team developed a 'harness' that guides models toward high-quality task decomposition during training, with subagents constrained to return properly formatted results; the resulting trajectories are then used as supervised fine-tuning data. Their resulting model, SearchSwarm-30B-A3B, achieves scores of 68.1 on BrowseComp and 73.3 on BrowseComp-ZH, which the authors report as the best results among models of comparable scale. The researchers note that training data for delegation behavior is scarce in naturally occurring text and that this capability has been largely unexplored in the open-source community. The team plans to release the harness, model weights, and training data to support further research.
What's missing
The paper does not report results on benchmarks beyond BrowseComp and BrowseComp-ZH, leaving open how SearchSwarm performs on broader agentic or general reasoning evaluations. Key limitations — such as failure modes of the delegation harness, sensitivity to subagent quality, and scalability beyond the 30B parameter range — are not detailed in the abstract. It is also unclear how the model compares to closed-source frontier models on the same benchmarks.
What different sources said
- arXiv cs.AICenter
SearchSwarm: Towards Delegation Intelligence in Agentic LLMs for Long-Horizon Deep Research
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.