Squeeze Evolve: Unified Multi-Model Orchestration for Verifier-Free Evolution
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917398918987776 |
|---|---|
| author | Maheswaran, Monishwaran Lakhani, Leon Zhou, Zhongzhu Yang, Shijia Wang, Junxiong Hooper, Coleman Hu, Yuezhou Tiwari, Rishabh Wang, Jue Singh, Harman Wu, Qingyang Jian, Yuqing Zhang, Ce Keutzer, Kurt Dao, Tri Wu, Xiaoxia Athiwaratkun, Ben Zou, James Xu, Chenfeng |
| author_facet | Maheswaran, Monishwaran Lakhani, Leon Zhou, Zhongzhu Yang, Shijia Wang, Junxiong Hooper, Coleman Hu, Yuezhou Tiwari, Rishabh Wang, Jue Singh, Harman Wu, Qingyang Jian, Yuqing Zhang, Ce Keutzer, Kurt Dao, Tri Wu, Xiaoxia Athiwaratkun, Ben Zou, James Xu, Chenfeng |
| contents | We show that verifier-free evolution is bottlenecked by both diversity and efficiency: without external correction, repeated evolution accelerates collapse toward narrow modes, while the uniform use of a high-cost model wastes compute and quickly becomes economically impractical. We introduce Squeeze Evolve, a unified multi-model orchestration framework for verifier-free evolutionary inference. Our approach is guided by a simple principle: allocate model capability where it has the highest marginal utility. Stronger models are reserved for high-impact stages, while cheaper models handle the other stages at much lower costs. This principle addresses diversity and cost-efficiency jointly while remaining lightweight. Squeeze Evolve naturally supports open-source, closed-source, and mixed-model deployments. Across AIME 2025, HMMT 2025, LiveCodeBench V6, GPQA-Diamond, ARC-AGI-V2, and multimodal vision benchmarks, such as MMMU-Pro and BabyVision, Squeeze Evolve consistently improves the cost-capability frontier over single-model evolution and achieves new state-of-the-art results on several tasks. Empirically, Squeeze Evolve reduces API cost by up to $\sim$3$\times$ and increases fixed-budget serving throughput by up to $\sim$10$\times$. Moreover, on discovery tasks, Squeeze Evolve is the first verifier-free evolutionary method to match, and in some cases exceed, the performance of verifier-based evolutionary methods. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_07725 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Squeeze Evolve: Unified Multi-Model Orchestration for Verifier-Free Evolution Maheswaran, Monishwaran Lakhani, Leon Zhou, Zhongzhu Yang, Shijia Wang, Junxiong Hooper, Coleman Hu, Yuezhou Tiwari, Rishabh Wang, Jue Singh, Harman Wu, Qingyang Jian, Yuqing Zhang, Ce Keutzer, Kurt Dao, Tri Wu, Xiaoxia Athiwaratkun, Ben Zou, James Xu, Chenfeng Artificial Intelligence Computation and Language We show that verifier-free evolution is bottlenecked by both diversity and efficiency: without external correction, repeated evolution accelerates collapse toward narrow modes, while the uniform use of a high-cost model wastes compute and quickly becomes economically impractical. We introduce Squeeze Evolve, a unified multi-model orchestration framework for verifier-free evolutionary inference. Our approach is guided by a simple principle: allocate model capability where it has the highest marginal utility. Stronger models are reserved for high-impact stages, while cheaper models handle the other stages at much lower costs. This principle addresses diversity and cost-efficiency jointly while remaining lightweight. Squeeze Evolve naturally supports open-source, closed-source, and mixed-model deployments. Across AIME 2025, HMMT 2025, LiveCodeBench V6, GPQA-Diamond, ARC-AGI-V2, and multimodal vision benchmarks, such as MMMU-Pro and BabyVision, Squeeze Evolve consistently improves the cost-capability frontier over single-model evolution and achieves new state-of-the-art results on several tasks. Empirically, Squeeze Evolve reduces API cost by up to $\sim$3$\times$ and increases fixed-budget serving throughput by up to $\sim$10$\times$. Moreover, on discovery tasks, Squeeze Evolve is the first verifier-free evolutionary method to match, and in some cases exceed, the performance of verifier-based evolutionary methods. |
| title | Squeeze Evolve: Unified Multi-Model Orchestration for Verifier-Free Evolution |
| topic | Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2604.07725 |