AMix-1: A Pathway to Test-Time Scalable Protein Foundation Model

2025/07/11 by Changze Lv, Jiang Zhou, Lv, Changze +47 · 1 voice · 1 citation
Biochemistry, Genetics and Molecular Biology · Computer Science · #Artificial Intelligence (cs.AI) #Biomolecules (q-bio.BM) #FOS: Biological sciences #FOS: Computer and information sciences #cs.AI #q-bio.BM

paper · pdf · doi:10.48550/arxiv.2507.08920

arxiv published 2025/07/11 · arxiv updated 2026/06/05

Abstract

We introduce AMix-1, a powerful protein foundation model built on Bayesian Flow Networks and empowered by a systematic training methodology, encompassing pretraining scaling laws, emergent capability analysis, in-context learning mechanism, and test-time scaling algorithm. To guarantee robust scalability, we establish a predictive scaling law and reveal the progressive emergence of structural understanding via loss perspective, culminating in a strong 1.7-billion model. Building on this foundation, we devise a multiple sequence alignment (MSA)-based in-context learning strategy to unify protein design into a general framework, where AMix-1 recognizes deep evolutionary signals among MSAs and consistently generates structurally and functionally coherent proteins. This framework enables the successful design of a dramatically improved AmeR variant with an up to 50× activity increase over its wild type. Pushing the boundaries of protein engineering, we further empower AMix-1 with an evolutionary test-time scaling algorithm for in silico directed evolution that delivers substantial, scalable performance gains as verification budgets are intensified, laying the groundwork for next-generation lab-in-the-loop protein design.

Citations

• Steering Protein Family Design through Profile Bayesian Flow
• s1: Simple test-time scaling
• Inference-Time Scaling for Diffusion Models beyond Scaling Denoising Steps
• Scaling Properties of Diffusion Models for Perceptual Tasks
• Scaling Laws For Diffusion Transformers
• Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters
• Large Language Monkeys: Scaling Inference Compute with Repeated Sampling
• Accurate structure prediction of biomolecular interactions with AlphaFold 3
• MolCRAFT: Structure-Based Drug Design in Continuous Parameter Space
• Understanding Emergent Abilities of Language Models from the Loss Perspective
• Diffusion Language Models Are Versatile Protein Learners
• CLIPZyme: Reaction-Conditioned Virtual Screening of Enzymes
• Bayesian Flow Networks
• GPT-4 Technical Report
• Evolutionary-scale prediction of atomic-level protein structure with a language model
• Scaling Instruction-Finetuned Language Models
• GLM-130B: An Open Bilingual Pre-trained Model
• Learning inverse folding from millions of predicted structures
• Emergent Abilities of Large Language Models
• Tranception: protein fitness prediction with autoregressive transformers and inference-time retrieval
• PaLM: Scaling Language Modeling with Pathways
• BNAI, NO-TOKEN, and MIND-UNITY: Pillars of a Systemic Revolution in Artificial Intelligence
• КОМПЬЮТЕРНАЯ ТОМОГРАФИЯ – ОСНОВА ОПТИМИЗАЦИИ МИНИТОРАКОТОМНОГО ДОСТУПА
• Highly accurate protein structure prediction with AlphaFold
• RoFormer: Enhanced Transformer with Rotary Position Embedding
• Language Models are Few-Shot Learners
• Improved protein structure prediction using potentials from deep learning
• Attention Is All You Need
• Inference Scaling Laws: An Empirical Analysis of Compute-Optimal Inference for Problem-Solving with Language Models

Cited by

• Evolutionary Profiles for Protein Fitness Prediction

Discussions

• Changze Lv, et al.: AMix-1: A Pathway to Test-Time Scalable Protein Foundation Model https://arxiv.org/abs/2507.08920 https://arxiv.org/pdf/2507.08920 https://arxiv.org/html/2507.08920 [bsky, 0 points, 0 comments]

Related

• AMix-2: Establishing Protein as a Native Modality in Large Language Models
• Adaptive Protein Design Protocols and Middleware
• TadA-Bench: A Million-Variant Benchmark for Future-Round Discovery Toward Agentic Protein Engineering
• Proteina: Scaling Flow-based Protein Structure Generative Models
• One protein is all you need