# Frontier LLMs are effective batch optimizers: Assessing reasoning models in continuous and discrete settings

arXiv: https://arxiv.org/abs/2609.03177 · HTML: https://mlsanity.com/paper/2609.03177/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 23 of its 39 references that are in the map.
- First posted on arXiv 2 Sep 2026.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 23 of 23)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Large Language Models as Optimizers](https://mlsanity.com/paper/2309.03409/) (2023)
- [A Tutorial on Bayesian Optimization](https://mlsanity.com/paper/1807.02811/) (2018)
- [Sample Efficiency Matters: A Benchmark for Practical Molecular Optimization](https://mlsanity.com/paper/2206.12411/) (2022)
- [Large Language Models to Enhance Bayesian Optimization](https://mlsanity.com/paper/2402.03921/) (2024)
- [ChemBERTa-2: Towards Chemical Foundation Models](https://mlsanity.com/paper/2209.01712/) (2022)
- [Efficient Evolutionary Search Over Chemical Space with Large Language Models](https://mlsanity.com/paper/2406.16976/) (2024)
- [Training a Scientific Reasoning Model for Chemistry](https://mlsanity.com/paper/2506.17238/) (2025)
- [GenMol: A Drug Discovery Generalist with Discrete Diffusion](https://mlsanity.com/paper/2501.06158/) (2025)
- [Gotta be SAFE: A New Framework for Molecular Design](https://mlsanity.com/paper/2310.10773/) (2023)
- [Amortized Tree Generation for Bottom-up Synthesis Planning and Synthesizable Molecular Design](https://mlsanity.com/paper/2110.06389/) (2021)
- [LICO: Large Language Models for In-Context Molecular Optimization](https://mlsanity.com/paper/2406.18851/) (2024)
- [Bayesian Optimization of Catalysis With In-Context Learning](https://mlsanity.com/paper/2304.05341/) (2023)
- [Composer 2 Technical Report](https://mlsanity.com/paper/2603.24477/) (2026)
- [Exploring the True Potential: Evaluating the Black-box Optimization Capability of Large Language Models](https://mlsanity.com/paper/2404.06290/) (2024)
- [Reinforced In-Context Black-Box Optimization](https://mlsanity.com/paper/2402.17423/) (2024)
- [Large language models as uncertainty-calibrated optimizers for experimental discovery](https://mlsanity.com/paper/2504.06265/) (2025)
- [Funneled Bayesian Optimization for Design, Tuning and Control of Autonomous Systems](https://mlsanity.com/paper/1610.00366/) (2016)
- [ExLLM: Experience-Enhanced LLM Optimization for Molecular Design and Beyond](https://mlsanity.com/paper/2502.12845/) (2025)
- [Evaluating the Progression of Large Language Model Capabilities for Small-Molecule Drug Design](https://mlsanity.com/paper/2604.16279/) (2026)
- [GP-MoLFormer-Sim: Test Time Molecular Optimization through Contextual Similarity Guidance](https://mlsanity.com/paper/2506.05628/) (2025)
- [BioBO: Biology-informed Bayesian Optimization for Perturbation Design](https://mlsanity.com/paper/2509.19988/) (2025)
- [Sample Efficient Generative Optimization for Molecular Design](https://mlsanity.com/paper/2607.12488/) (2026)

## Abstract (arXiv)

Frontier large language models (LLMs) have become attractive priors for optimization due to their large-scale pretraining that enables them to navigate a variety of optimization settings. However, the effectiveness of modern reasoning LLMs in batch optimization settings remains underexplored. Here we investigate the performance of the current generation of frontier LLMs as batch optimizers in both continuous and discrete settings. We find that while LLMs are competitive zero-shot batch optimizers for numerical test functions, their performance is brittle compared to classical non-LLM optimization approaches. However, LLM priors are significantly better in semantically rich settings, indicating that their batch optimization behavior is highly effective when navigating and reasoning over the discrete spaces most similar in structure to their pretraining data.