A Closer Look at Agentic BBO: Benchmarking LLM Agents for Black-Box Optimization

Ming Chen; Rong-Xi Tan; Ke Xue; Yu-Jie Zhou; Taiye Lu; Zhi-Xuan Gao; Peng Xie; Zijun Shen

cs.LGOct 8, 2026

A Closer Look at Agentic BBO: Benchmarking LLM Agents for Black-Box Optimization

Authors: Ming Chen, Rong-Xi Tan, Ke Xue, Yu-Jie Zhou, Taiye Lu, Zhi-Xuan Gao, Peng Xie, Zijun Shen, +3 more

Organizations: State Key Laboratory of Novel Software Technology, Nanjing University · School of Artificial Intelligence, Nanjing University

Abstract

Black-box optimization (BBO) arises in many scientific and engineering problems where objective evaluations are expensive and limited. Recent large language model (LLM) agents offer a new way to approach BBO by combining task semantics, computation, optimization tools, and feedback-driven decision making, showing great potential due to the integration with mathematically rigorous tools. However, existing agentic BBO studies use different task domains and system configurations, making their results difficult to compare and the effects of individual design choices hard to isolate. We therefore introduce AgenticBBO-Bench, a cross-domain benchmark for agentic BBO spanning synthetic functions, hyperparameter optimization, database tuning, chip design, and molecular design under a unified finite-budget evaluation protocol. In our experiments, agentic BBO achieves higher family-averaged scores than direct LLM-based methods in all five domains and outperforms the best numerical optimizers in four. We further study three factors shaping agent performance: optimization tools, task information and prior knowledge, and the role of the LLM during search. Our results show that additional numerical tools do not consistently improve performance, task semantics are broadly useful while more specific priors are less reliable, and numerical optimizers can effectively absorb gains from search trajectories established by the agent. Finally, we introduce a five-task frontier challenge within AgenticBBO-Bench and evaluate seven LLMs under the Codex agent harness, where GPT-6 Astra and DeepSeek-V4.1-Flash lie on the Pareto frontier of performance and cost among the evaluated models. Our code is available at https://github.com/lamda-bbo/agentic-bbo.

Figures & tables

Figure 1: Frontier-model comparison on AgenticBBO-Bench. Left: overall benchmark scores. Right: score versus official list-price token cost per task (log scale). Models with similar optimization performance can incur substantially different inference costs.

Figure 2: Comparison between traditional Bayesian optimization and Agentic BBO. Traditional BO follows a fixed optimizer-driven procedure, whereas Agentic BBO lets an LLM agent organize the intermediate decision process before each expensive evaluation. The actions shown in the agent loop are representative examples rather than a fixed or exhaustive workflow: the agent may dynamically choose, skip, repeat, combine, or introduce other actions before proposing a candidate.

Family	Search Space	Semantics	Feasibility Constraints	Evaluator
Synthetic	Continuous	–	–	Analytic function
HPO	Mixed	$\checkmark$	–	Model training
Database Tuning	Mixed	$\checkmark$	–	Learned surrogate
Chip Design	Continuous	$\checkmark$	$\checkmark$	Placement evaluator
Molecular Design	Structured discrete	$\checkmark$	$\checkmark$	Molecular scoring

Table 1: Overview of the five benchmark task families. Semantics indicates whether task or variable meanings are exposed to the optimizer, while Feasibility Constraints indicates whether the search space contains candidates that may be infeasible or invalid. Evaluator describes how objective values are obtained in each task family.

Table 2: Results of the broad benchmark and numerical-tool study. Scores are normalized such that higher values indicate better optimization performance, and the best result in each comparison is shown in bold . A dash indicates that the method is not applicable to that task family.

Panel of Figure 3 — Figure 3: Tool use in Agentic BBO. (a) Tool-use patterns across the diagnostic tasks. (b) A representative Bigblue1 trajectory in which the agent fits a proxy model and uses it to refine the placement.

Panel of Figure 4 — Figure 4: Effect of task information and prior knowledge. (a) Performance on six real-world tasks with anonymous inputs, task semantics, or additional domain priors. (b) Performance on controlled objectives under different forms of prior information.

Table 3: Results of the LLM-role study and the five-task frontier challenge. Higher scores indicate better optimization performance, and the best result in each comparison is shown in bold .

Appendix figures & tables3 assets

Supplementary material from the paper’s appendix.

Appendix

Experimental panel	Tasks	Initial + new evaluations
Broad BBOB	24	$20+100$
Broad HPO	25	$5+25$
Broad DBTune	6	$50+200$
Broad BBOPlace	12	$50+200$
Broad GuacaMol	10	$50+200$
Tool / policy diagnostics: BBOB	2	$20+50$

Appendix

Table 4: Evaluation budgets used across the experimental panels. “Initial + new” denotes the number of shared initial observations followed by the number of additional objective evaluations available to each method.

Session	v0	v1	v2	v3	v4	v5	v6	Pick	Held-out $S$
BBOB 1	0.472	0.654	0.660	0.651	0.636	0.634	0.612	v5	0.548
BBOB 2	0.472	0.470	0.417	0.442	0.488	0.460	0.393	v5	0.484
BBOB 3	0.472	0.670	0.683	0.759	0.735	0.741	0.700	v4	0.630
HPO 1	0.507	0.574	0.577	0.607	0.594	0.594	0.594	v4	0.532
HPO 2	0.507	0.285	0.603	0.625	0.589	0.622	0.615	v5	0.531
HPO 3	0.507	0.577	0.492	0.455	0.290	0.597	0.607	v6	0.279

Appendix

Table 5: Development trajectories and held-out performance. Development entries average four tasks and two seeds, while held-out

S

averages two disjoint tasks and four seeds. Bold marks the historically selected version.

Session	Pick	Implemented search logic
BBOB 1	v5	GP-LogEI with an adaptive incumbent-centered candidate region.
BBOB 2	v5	UCB for the first 12 new evaluations, followed by EI on the same GP.
BBOB 3	v4	Matérn-5/2 GP with LogEI, conditional inverse-hyperbolic-sine response compression, and a trust region.
HPO 1	v4	Warped-coordinate GP with global EI candidate pools and dimension-dependent greedy local search.
HPO 2	v5	Warped-coordinate GP-EI with an adaptive trust region and separation from observed candidates.
HPO 3	v6	Reference GP-EI augmented with scheduled Halton coverage probes and a structured-pool duplicate fallback.

Appendix

Table 6: Search logic implemented by the six selected programs.

Explore similar work

Agentic Bayesian Optimization through Surrogate-Augmented Autoresearch
Jul 31, 2026Paul Brunzema, Louis Tiao, Nhat Le +3Surrogate-Assisted Optimization Bayesian Optimization
Meta · RWTH Aachen University, Germany
BoLT: A Benchmark to Democratize Black-box Optimization Research for Expensive LLM Tasks
May 16, 2026Ruth Wan Theng Chew, Zhiliang Chen, Apivich Hemachandra +1Bayesian Optimization Benchmark Design
National University of Singapore
Bayesian Optimization with Rich Auxiliary Information via LLMs
Sep 16, 2026Tejus Gupta, Efe Mert Karagözlü, Rohit Sonker +2Bayesian Optimization Large Language Model-Guided Optimization
School of Computer Science Carnegie Mellon University Pittsburgh, PA 15213

Jul 31, 2026, cs.LGJ/K move · Enter open · S save

Agentic Bayesian Optimization through Surrogate-Augmented Autoresearch

Paul Brunzema, Louis Tiao, Nhat Le, Kevin De Angeli+2

Meta · RWTH Aachen University, Germany

Surrogate-Assisted Optimization Bayesian Optimization Large Language Model-Guided Optimization

Open paper AskPDF arXiv

Bayesian optimization (BO) has become the standard tool for sample-efficient optimization and owes its efficiency to uncertainty-aware search driven by generic statistical priors. Richer domain priors can improve BO in principle, but encoding them through tailored kernels or problem structure is difficult and rarely done in practice. LLMs can help sidestep this difficulty by making informal priors from natural language, code, and documentation directly available to the optimizer. However, existing LLM-based BO methods either insert the LLM into a fixed role (surrogate, acquisition proxy, or configuration interface) or hand it broad control, sacrificing the systematic exploration that makes BO reliable. We introduce agentic Bayesian optimization: a paradigm in which an LLM agent is the central decision maker in the BO loop while a Bayesian backend provides the uncertainty-aware optimization substrate. The agent configures the problem, queries the backend, selects and commits evaluations, and can revise the optimization strategy during the run by tightening bounds, switching acquisition functions, proposing targeted evaluations, or even reframing the problem following new instructions or observed evidence. We instantiate this idea in Sara, a surrogate-augmented autoresearch agent, and lenz, a modular BoTorch-based backend that the agent can inspect and modify through a structured interface. Across synthetic and real-world benchmarks, Sara preserves the reliability of state-of-the-art BO without prior knowledge, outperforms LLM-based baselines, and uses natural-language priors to improve beyond standard BO. We further demonstrate the practical value of agentic BO in dynamic settings, where Sara reconfigures the full optimization problem on the fly as requirements change, a capability not previously available in standard BO.

Abstract

Explore similar work

Agentic Bayesian Optimization through Surrogate-Augmented Autoresearch

BoLT: A Benchmark to Democratize Black-box Optimization Research for Expensive LLM Tasks

Bayesian Optimization with Rich Auxiliary Information via LLMs