check chemical validity

Existing benchmarks for chemical reasoning evaluation focus solely on final answers, making it difficult to detect logical errors in intermediate reasoning steps. To address this limitation, this work proposes ChemCoTBench-V2, a novel benchmark that employs expert-designed structured templates to guide models in generating verifiable intermediate reasoning states. By integrating deterministic chemical rules, reference trajectory alignment, and oracle-verifiable state constraints, the framework enables low-cost, auditable process-level evaluation without requiring human or LLM-based adjudication. This approach is the first to support state-constraint verification and precise error localization in open-ended tasks, revealing a significant discrepancy between answer correctness and reasoning consistency across mainstream large language models. It further facilitates fine-grained model comparison and identification of the first erroneous step in reasoning trajectories.

This work addresses the lack of systematic evaluation for symbolic, verifiable reasoning over molecular graph structures in current chemical large language models. Existing benchmarks often suffer from label bias or information leakage, hindering precise diagnosis of model shortcomings. To bridge this gap, we propose MolecularIQ—the first evaluation framework specifically designed for symbolic reasoning on molecular graphs. By integrating molecular graph representations, symbolic logic verification, and carefully structured reasoning tasks, MolecularIQ establishes a fine-grained benchmark that effectively uncovers systematic failure modes of contemporary models across specific molecular structures and reasoning challenges. This framework provides interpretable diagnostic insights and actionable directions for developing chemical large language models with faithful structural understanding capabilities.

Current large language models (LLMs) lack systematic evaluation of structured, constraint-aware reasoning capabilities in chemistry—particularly for molecular property optimization and reaction prediction. Method: We propose ChemCoTBench, the first benchmark framework for “slow-thinking” chemical reasoning, introducing a novel “modular chemical operation” paradigm (addition, deletion, substitution) that formalizes molecular transformations as interpretable, stepwise symbolic processes. It integrates graph-structured molecular representation, constraint-driven chain-of-thought (CoT) reasoning, and a manually curated dataset. Contribution/Results: ChemCoTBench establishes an evaluation framework measuring both reasoning-path traceability and rule adherence. Experiments demonstrate substantial improvements in LLMs’ ability to model domain-specific constraints and reaction rules, advancing trustworthy, interpretable chemical AI.

本文提出ChemXRG框架，通过生成可执行程序解决3D化学反应动画的可视化问题，利用0.8B模型生成领域特定语言并验证转换。

This work aims to achieve computable chemical synthesis—i.e., precise, code-driven control of reaction pathways on general-purpose reconfigurable hardware to automate the synthesis of any stable, isolable molecule while satisfying mass conservation, finite reaction time, and analytical detectability constraints. Method: We introduce the “chemputation” paradigm, modeling synthesis as graph transformations over the space Reagents × Process × Catalyst. We formally define and prove the Universal Chemical Synthesis Theorem, introduce the notion of “analytically reachable quantity” for molecules, and establish dynamic error correction as essential. Our end-to-end implementation integrates the Chemputer hardware platform, the chempiler compiler, assembly-theory–driven reachability analysis, and a real-time sensing feedback framework. Results: Experimental validation demonstrates that chemical reactions are intrinsically programmable, observable, and correctable graph operations. We identify reactor count and sensor bandwidth as critical scalability bottlenecks for chemputation.

This work addresses the fragmented and labor-intensive pipeline in scientific machine learning—from data acquisition to model deployment—by proposing an end-to-end, high-throughput, and agent-collaborative reproducible workflow platform. The platform enables full automation of data collection, processing, modeling, validation, selection, and reporting through modular components, configuration-driven mechanisms, and standardized artifacts. Key technical features include plug-and-play model architectures, deterministic data splitting, batch execution, and structured outputs. Evaluated on quantum mechanics, physicochemical property, and bioactivity prediction tasks, the system achieves state-of-the-art performance and successfully generalizes to non-molecular domains such as time series, significantly enhancing research efficiency and reproducibility.

This work addresses the challenge of accurately translating natural language instructions into chemically valid molecular structures under stringent constraints in AI-driven drug discovery. The authors propose Mol-Debate, a novel framework that introduces a multi-agent debate mechanism to simulate the multi-perspective critique and iterative refinement characteristic of real-world drug design. Through a generate–debate–optimize loop, Mol-Debate uniquely coordinates developer and evaluator roles, integrating global and local structural reasoning with both static and dynamic chemical knowledge. Experimental results demonstrate that Mol-Debate achieves a 59.82% exact match rate on ChEBI-20 and a 50.52% weighted success rate on S²-Bench, substantially outperforming existing baselines.

This study addresses the unclear performance bottlenecks of AI agents in drug discovery by proposing MAGI, a modular agent designed to investigate whether tool orchestration or predictive model accuracy constrains practical outcomes. Methodologically, MAGI employs an open modular architecture coordinating molecular design, optimization monitoring, and SAR analysis with self-revising strategies. It integrates a dual-pathway mechanism combining direct large language model generation with REINFORT-delegated generation, alongside a pluggable scoring service contract. Validation on retrospective pharmaceutical projects reveals that the primary bottleneck lies in the applicability domain of scoring models rather than agent orchestration capabilities. Furthermore, MAGI generates molecules approaching expert-level quality and integrates effectively into existing computational chemistry workflows, thereby clarifying the practical role of AI agents in real-world drug discovery scenarios.

This study addresses the challenge that large language models (LLMs) face in handling quantitatively constrained molecular modifications and systematic revisions within combinatorial chemistry. To this end, we propose TMCS, a framework that formalizes chemical problem-solving as an interpretable, tool-augmented workflow. By leveraging multi-agent collaboration, TMCS unifies the stages of molecular generation, understanding, and editing. Furthermore, it incorporates few-shot trajectory memory and structured reflection mechanisms to enable closed-loop iterative optimization. Experimental results demonstrate that TMCS substantially enhances model performance across diverse chemical reasoning tasks, achieving state-of-the-art results on both open-source and proprietary LLMs.

This study addresses the limited cross-source generalization and misalignment with expert decision-making in AI-generated chemical reaction validators. We construct a multi-source reaction benchmark comprising 751 expert-annotated instances and systematically evaluate the verification performance of large language models and forward prediction models. By employing diverse negative sampling techniques for cross-analysis, we expose the generalization deficiencies of existing validators and release a corpus containing tens of millions of negative samples. Experimental results demonstrate that no single validator consistently outperforms others across all sources. Furthermore, while training on mixed negative samples improves overall AUROC, such gains fail to transfer effectively to model-proposed scenarios.