RetroPlanner: Multi-Route Expansion and Multi-Objective Reasoning for LLM-Based Retrosynthesis Planning

Ye Rin Kim; Yongjun Jeong; Seongjun Yun; Hanbum Ko; Soohyun Oh; Sungbin Lim; Sungwoong Kim

cs.AIOct 12, 2026

RetroPlanner: Multi-Route Expansion and Multi-Objective Reasoning for LLM-Based Retrosynthesis Planning

Authors: Ye Rin Kim, Yongjun Jeong, Seongjun Yun, Hanbum Ko, Soohyun Oh, Sungbin Lim, Sungwoong Kim

Organizations: Korea University, Department of Artificial Intelligence · Korea University, Department of Statistics

Abstract

Multi-step retrosynthesis planning is often considered solved once a route to purchasable starting materials is found, but such a route should also be practical in real-world chemistry. Recent LLM-based planners formulate this process as sequential decision-making over an AND-OR graph, yet typically expand a single molecule per turn and are optimized for solvability alone, leaving alternative routes uncompared across the multiple objectives that determine practicality before the search budget is committed. We introduce RetroPlanner, an LLM-based planner that jointly compares multiple routes to steer the search toward higher synthesizability, as measured by reaction plausibility, round-trip consistency, and starting-material price. RetroPlanner is trained to reason over chemical evidence: structural progress from the target molecule, bond changes and reaction classes, where candidate routes diverge from one another, and how they rank across multiple objectives. Experiments on multi-step retrosynthesis benchmarks demonstrate that RetroPlanner recovers more alternative routes than prior LLM-based planners while improving synthetic route quality and maintaining a high success rate.

Figures & tables

Figure 1: Since routes that dominate the agent’s returned routes on reaction plausibility, round-trip consistency, and starting-material price are often already exposed during search but left unexpanded, RetroPlanner recovers them by jointly expanding multiple frontier molecules and comparing them in a single multi-objective decision per turn.

Figure 2: Planning on the same target under three expansion patterns, corresponding to the strategies analyzed in Section 4.1 . (a) Single-route expansion extends one route per turn, as in Retro-R1. (b) Sequential multi-route expansion grows several routes, but each turn opens and ranks one frontier molecule, as in RetroAgent. (c) Parallel multi-route expansion, used by RetroPlanner , runs three stages per turn: open selects multiple molecules from the frontier set

\mathcal{O}_{t}

to expand at once, rank jointly orders the reaction candidates of the opened molecules for expansion through multi-objective reasoning over reaction plausibility (P), round-trip consistency (RT), and starting-material price (C). done commits routes whose leaves are all in the building-block set

\mathcal{BB}

Figure 3: (a) Multi-objective reasoning. Reaction candidates are compared on three primary decision signals and auxiliary signals, which are referenced to order the candidates of each molecule. (b) Reasoning trace construction. Structural and reaction-level evidence is collected for the target and its candidates; a teacher LLM writes the MOLECULE / DISCONNECTIONS / DECIDING trace from that evidence, and each field is verified against the search graph, with detected inconsistencies returned as repair feedback until the trace passes.

USPTO-190
Method	Search Outcome		Route Quality
Method	Success Rate $\uparrow$	Routes $\uparrow$	Better-Than-Ref $\uparrow$	Plausibility $\uparrow$	Round-Trip $\uparrow$	Price $\downarrow$
Search Algorithms
MCTS	44.74%	16.88	27.89%	43.33%	25.88%	391.19
Retro*-0	52.63%	29.55	34.21%	49.16%	38.99%	328.02
Retro*	57.89%	33.34	38.95%	54.93%	41.64%	281.53

Table 1: Main results on multi-step retrosynthesis planning on USPTO-190 and ChEMBL-1000 with search budget

N_{B}=50

. Search Outcome measures whether and how many routes are found, Route Quality evaluates the found routes on reaction plausibility, round-trip consistency, and starting-material price. Better-Than-Ref is N/A on ChEMBL-1000, which provides no reference routes. Best and second-best results per dataset are shown in bold and underlined .

Figure 4: Overall search time versus molecules opened per rollout. Although the two multi-route arms open a comparable number of molecules per episode, Sequential spends more time per molecule than Parallel .

Variant	Schema			Rel.		Par.	Search Outcome				Diversity				Quality
Variant	M	D	C	div	ord	Par.	Succ. $\uparrow$		Routes $\uparrow$		1stCutDiff $\uparrow$		1stCuts $\uparrow$		BTR $\uparrow$
$L_{0}$	–	–	–	–	–	–	0.989		4.78		0.628		2.79		0.720
$L_{1}$	✓	–	–	–	–	$L_{0}$	0.989	( $+0.000$ )	5.49	( $+0.71$ )	0.691	( $+0.063$ )	3.15	( $+0.36$ )	0.769	( $+0.049$ )
$L_{2}$	✓	✓	–	–	–	$L_{1}$	0.995	( $+0.006$ )	5.53	( $+0.04$ )	0.722	( $+0.031$ )	3.16	( $+0.01$ )	0.742	( $-0.027$ )
$L_{3\text{-div}}$	✓	✓	✓	✓	–	$L_{2}$	0.989	( $-0.006$ )	5.78	( $+0.25$ )	0.726	( $+0.004$ )	3.34	( $+0.18$ )	0.720	( $-0.022$ )
$L_{3\text{-ord}}$	✓	✓	✓	–	✓	$L_{2}$	0.979	( $-0.016$ )	5.08	( $-0.45$ )	0.756	( $+0.034$ )	3.20	( $+0.04$ )	0.737	( $-0.005$ )

Table 3: Ablation on the reasoning schema on USPTO-190, with all variants run under the same search budget of at most

N_{B}=300

. Schema blocks: M = MOLECULE , D = DISCONNECTIONS , and C = DECIDING . The Rel. columns indicate the diverging from ( div ) and ordered by ( ord ) subfields of DECIDING . Par. denotes the parent variant from which each delta is computed. Best results are in bold .

Appendix figures & tables4 assets

Supplementary material from the paper’s appendix.

Appendix

Training
Hyperparameter	Value
Base model	GPT-OSS-20B
Train / val trajectories	3,142 / 64
Epochs	2
Steps	429
Batch size	4 (packed, $\sim$ 393K tokens)

Appendix

Table 6: Training and decoding hyperparameters for RetroPlanner .

USPTO-190
Method	Success Rate (%) at Search Budget $N_{B}$
Method	$\mathbf{N_{B}=50}$	$\mathbf{N_{B}=100}$	$\mathbf{N_{B}=200}$	$\mathbf{N_{B}=300}$	$\mathbf{N_{B}=400}$	$\mathbf{N_{B}=500}$
Search Algorithms
MCTS	44.74	52.11	58.95	67.89	71.58	72.11
Retro*-0	52.63	59.47	63.68	70.00	72.11	74.74
Retro*	57.89	64.21	71.58	79.47	80.53	84.74

Appendix

Table 7: Success rate (%) under increasing search budgets on USPTO-190 and ChEMBL-1000, with R-SMILES as the single-step model for all methods. The search budget

N_{B}

is the number of unique single-step model calls. Best and second-best results are in bold and underlined , respectively.

Figure 5: Left: Attention from action tokens to the route rows of the input search graph, under different contents of the reasoning field. Attention is highest when the field holds the instance’s own analysis, compared with no reasoning before or after training, meaningless text, or the analysis of another instance. Right: Attention from diverging from tokens to the rows of the search graph in two examples. Attention concentrates on the referenced route (orange) rather than the current branch (green).

Figure 6: From step-level decisions to route-level objectives for a single target. Left: Routes recovered by RetroPlanner on three route-level objectives: the fraction of steps with plausibility $p\geq 0.05$ , the fraction of steps with round-trip consistency $rt\leq 5$ , and the total starting-material price $c$ , where lower is better. Pareto-optimal routes are shown in red and dominated routes in gray. Right: A Pareto-optimal route (A) and a dominated route (B) for the same target. Each reaction is annotated with its $p$ , $rt$ , and building-block price, and $\blacktriangle$ marks the objective that separated the leading candidates at that step. In route A, choices justified by different objectives at each step combine into a route that is not dominated on any objective.

Retrosynthesis

See all 23 papers

RetroAgent: Harnessing LLMs to Search Over Structured Memory for Agentic Retrosynthesis Planning
Jul 16, 2026Yanqiao Zhu, Jingru Gan, Xiaoqi Sun +8LLM Agent Harnesses LLM Agent Planning
UCLA · MIT · Amazon +1
URSA: Chemistry-Aware Benchmark for Utilitarian Retrosynthesis Assessment
Jul 6, 2026Bogdan Zagribelnyy, Ivan Ilin, Nikita Bondarev +7LLM Evaluation LLM Planning
Insilico Medicine AI Limited, Masdar City, Abu Dhabi, UAE · Independent researcher · Insilico Medicine Canada Inc., Montreal, Quebec, Canada +1
RetroReasoner: A Reasoning LLM for Strategic Retrosynthesis Prediction
Mar 13, 2026Hanbum Ko, Chanhui Lee, Ye Rin Kim +4AI-Assisted Scientific Research Language Modeling
Department of Artificial Intelligence, Korea University · Materials Intelligence Lab, LG AI Research · Department of Statistics, Korea University

Jul 16, 2026J/K move · Enter open · S save

RetroAgent: Harnessing LLMs to Search Over Structured Memory for Agentic Retrosynthesis Planning

Yanqiao Zhu, Jingru Gan, Xiaoqi Sun, Fang Sun+7

UCLA · MIT · Amazon+1

LLM Agent Harnesses LLM Agent Planning Memory-Augmented Language Model Agents Retrosynthesis

Open paper AskPDF arXiv

Multi-step retrosynthesis planning seeks to decompose a target molecule into commercially available building blocks through a sequence of feasible reactions. The vast combinatorial search space makes this task challenging even for expert chemists. Traditional methods combine tree search with offline-trained value networks that score candidates in isolation, without reasoning about complete multi-step routes. Recent work leverages Large Language Models (LLMs) for this task, but relies on simple interfaces that limit exploration of the full search space. We introduce RetroAgent, an LLM agent that bridges symbolic search and neural reasoning through a harness with structured memory. Through memory and chemistry tools, the agent observes the full search state, including explored routes, available alternatives, and properties of intermediates, enabling informed decisions grounded in both global progress and domain knowledge. Experiments on in-distribution and out-of-distribution benchmarks demonstrate that RetroAgent delivers strong performance and generalization.

Abstract

Retrosynthesis

Explore similar work

RetroAgent: Harnessing LLMs to Search Over Structured Memory for Agentic Retrosynthesis Planning

URSA: Chemistry-Aware Benchmark for Utilitarian Retrosynthesis Assessment

RetroReasoner: A Reasoning LLM for Strategic Retrosynthesis Prediction