# Multi-Objective Bayesian Optimization for Model Merging

arXiv: https://arxiv.org/abs/2608.14264 · HTML: https://mlsanity.com/paper/2608.14264/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 20 of its 31 references that are in the map.
- First posted on arXiv 14 Aug 2026.

## Builds on (most cited first, 20 of 20)

- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time](https://mlsanity.com/paper/2203.05482/) (2022)
- [Editing Models with Task Arithmetic](https://mlsanity.com/paper/2212.04089/) (2022)
- [TIES-Merging: Resolving Interference When Merging Models](https://mlsanity.com/paper/2306.01708/) (2023)
- [A Tutorial on Bayesian Optimization](https://mlsanity.com/paper/1807.02811/) (2018)
- [Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch](https://mlsanity.com/paper/2311.03099/) (2023)
- [Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities](https://mlsanity.com/paper/2408.07666/) (2024)
- [Arcee's MergeKit: A Toolkit for Merging Large Language Models](https://mlsanity.com/paper/2403.13257/) (2024)
- [Differentiable Expected Hypervolume Improvement for Parallel Multi-Objective Bayesian Optimization](https://mlsanity.com/paper/2006.05078/) (2020)
- [High-Dimensional Bayesian Optimization with Sparse Axis-Aligned Subspaces](https://mlsanity.com/paper/2103.00349/) (2021)
- [Parallel Bayesian Optimization of Multiple Noisy Objectives with Expected Hypervolume Improvement](https://mlsanity.com/paper/2105.08195/) (2021)
- [Practical Multi-fidelity Bayesian Optimization for Hyperparameter Tuning](https://mlsanity.com/paper/1903.04703/) (2019)
- [Parameter Competition Balancing for Model Merging](https://mlsanity.com/paper/2410.02396/) (2024)
- [Checkpoint Merging via Bayesian Optimization in LLM Pretraining](https://mlsanity.com/paper/2403.19390/) (2024)
- [MAP: Low-compute Model Merging with Amortized Pareto Fronts via Quadratic Approximation](https://mlsanity.com/paper/2406.07529/) (2024)
- [It's Morphing Time: Unleashing the Potential of Multiple LLMs via Multi-objective Optimization](https://mlsanity.com/paper/2407.00487/) (2024)
- [BECAME: BayEsian Continual Learning with Adaptive Model MErging](https://mlsanity.com/paper/2504.02666/) (2025)
- [Model Fusion through Bayesian Optimization in Language Model Fine-Tuning](https://mlsanity.com/paper/2411.06710/) (2024)
- [Model Merging is Secretly Certifiable: Non-Vacuous Generalisation Bounds for Low-Shot Learning](https://mlsanity.com/paper/2505.15798/) (2025)

## Abstract (arXiv)

Model merging combines trained models directly in weight space, offering a compute-efficient alternative to additional fine-tuning. Selecting merge parameters is nevertheless difficult because downstream evaluations are expensive, gradients are unavailable, and source capabilities can conflict. We formulate merge-parameter selection as a black-box multi-objective optimization problem and introduce MOBO-Merge, a merge-operator agnostic framework that uses multi-objective Bayesian optimization to approximate the Pareto front under a limited evaluation budget. We evaluate Qwen3-4B and Llama-3.1-8B in two-model instruction-math and three-model instruction-math-code settings using Linear, SLERP, TIES, and block-wise merge operators. On held-out benchmark partitions, MOBO-Merge obtains higher mean hypervolume than random search in 11 of 12 reported comparisons. The gain is small for one-dimensional Linear interpolation but substantially larger for several TIES, block-wise, and three-objective searches. No merge operator is uniformly best: TIES leads in three of four family-setting combinations, whereas Block-Linear 4x is strongest for the Llama three-model merge. These results show that multi-objective Bayesian optimization is valuable as a search layer for expressive merge parameterizations.