---
title: "SciReasoner: A Multimodal Scientific Foundation Model for Native Structural Reasoning Across Proteins, Small Molecules and Inorganic Crystals"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2607.07708v1"
permalink: "https://feed.astralor.com/items/itm_6dcaed6fe13a022b31dee10e"
published: "2026-07-08T17:59:59.000Z"
score: 81
channel: "ai"
tags: ["SciReasoner","Explainable AI","Multimodal Scientific Foundation Model","Structural Reasoning","Interdisciplinary"]
---

# SciReasoner: A Multimodal Scientific Foundation Model for Native Structural Reasoning Across Proteins, Small Molecules and Inorganic Crystals

## AI commentary

> This work really stands out to me—it makes a genuine attempt to address the 'black box' pain point of AI in science by turning structural tokens into reasonable evidence, combining accuracy gains with interpretability. For researchers in computational chemistry, structural biology, or materials genomics, SciReasoner's interdisciplinary capability and expert-validation results are an important reference. Notably, the authors have open-sourced the model and reasoning traces, making the ecosystem development worth watching.

## At a glance

This paper introduces SciReasoner, a multimodal scientific foundation model that performs native structural reasoning across proteins, small molecules, and inorganic crystals. By discretizing coordinates, topologies, and periodic connectivities into a unified structure-aware vocabulary, it treats structural tokens as addressable evidence units during reasoning. Across interdisciplinary benchmarks, SciReasoner achieves state-of-the-art performance on 67 out of 86 tasks—for example, improving Gene Ontology Cellular Component annotation Fmax from 0.42 to 0.55 for low-homology proteins, and raising single-step retrosynthesis accuracy from 0.63 to 0.72. Double-blind expert evaluations show its reasoning traces are preferred or comparable to a frontier large language model in 98% of cases.

## Key points

- SciReasoner converts spatial coordinates, topologies, and periodic connectivities into a unified structure-aware discrete vocabulary, enabling the model to reason with structure as addressable evidence.
- Across 86 benchmarks in biology, chemistry, and materials science, SciReasoner achieves state-of-the-art on 67 tasks, covering Gene Ontology prediction, retrosynthesis analysis, and band-gap classification.
- Double-blind expert evaluations show SciReasoner's reasoning traces are preferred or comparable to a frontier LLM in 98% of cases.

## Full text

Structure-property relationships are foundational to biology, chemistry and materials science, where function, reactivity and physical response emerge from spatial, chemical and periodic organization. Mechanistically explaining these relationships requires interpreting structural evidence through scientific principles and physical constraints, from stereochemistry and bonding to symmetry, energetics and periodic order. However, applying artificial intelligence to this process presents a joint challenge of representation and reasoning: models must preserve domain-native structural information while showing how specific evidence supports predictions under these constraints. Here we introduce SciReasoner, a multimodal scientific foundation model for native structural reasoning across proteins, small molecules and inorganic crystals. SciReasoner discretizes coordinates, topologies and periodic connectivities into a unified structure-aware vocabulary, treating structural tokens as addressable evidence units during reasoning. In homology-controlled Gene Ontology prediction, SciReasoner improves Cellular Component annotation for low-homology and orphan-like proteins, increasing $F_{\max}$ from 0.42 to 0.55. In chemistry, it raises single-step retrosynthesis accuracy from 0.63 to 0.72 while generating fragment-level disconnection and precursor-verification traces. In materials science, its representations separate elemental and compound phases and resolve high- and low-band-gap regimes. Across 86 benchmarks, SciReasoner achieves state-of-the-art performance on 67 tasks. Double-blind expert evaluation rates its reasoning traces as preferred or at least comparable to those of a frontier large language model in 98% of cases. By making structure an inspectable substrate for reasoning under scientific constraints, SciReasoner connects accurate prediction with interpretable scientific inference.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_6dcaed6fe13a022b31dee10e