Jun 5, 2026ScienceMaking Claude a chemist

2026-06-03 · Source: Anthropic Research · Field: Science & Research — Physical Sciences & Chemistry, Mathematics & Computational Sciences · Depth: Advanced, long

Summary

Anthropic is enhancing Claude's chemistry capabilities, specifically evaluating its performance on Nuclear Magnetic Resonance (NMR) spectrum analysis. A white paper details tests comparing Claude models (Opus 4.7, Opus 4.6, Sonnet 4.6) against dedicated software like ChemDraw and MestReNova on 20 compounds from synthetic chemistry preprints. For forward prediction, Opus 4.7 achieved an average hydrogen error of ±0.079 ppm, outperforming other tools and matching experimental splitting patterns more often. On carbon, Opus 4.7 and MestReNova were comparable at ±1.37 ppm and ±1.48 ppm, respectively. Crucially, Opus 4.7 successfully performed inverse prediction, elucidating structures from 1D NMR spectra and molecular formulas, a task typically left to chemists or requiring 2D NMR and specialized software. This demonstrates Claude's potential to assist with daily chemical translation and integration work.

Key takeaway

For synthetic chemists and research scientists grappling with time-consuming NMR analysis and structure elucidation, Claude Opus 4.7 offers a significant efficiency gain. You can now use a general-purpose AI to perform 1D NMR forward prediction with accuracy comparable to specialized software and uniquely tackle inverse prediction from spectra alone. Consider integrating Claude into your workflow for routine spectral interpretation and initial structure proposals to accelerate discovery.

Key insights

Claude Opus 4.7 demonstrates competitive accuracy in NMR spectrum prediction and novel capability in 1D NMR-based structure elucidation.

Principles

• Multimodal AI models can interpret diverse chemical representations.
• General-purpose LLMs can rival specialized chemistry software.
• AI can audit its reasoning step-by-step.

Method

Claude models (Opus 4.7, 4.6, Sonnet 4.6) were tested against ChemDraw and MestReNova on 20 compounds for forward NMR prediction and 15 for inverse structure elucidation using 1D NMR spectra and molecular formulas.

In practice

• Use Claude for translating between chemical representations.
• Apply Claude Opus 4.7 for 1D NMR structure elucidation.
• Integrate AI for routine data prediction in synthetic chemistry.

Topics

• NMR Spectroscopy
• Structure Elucidation
• Large Language Models
• Computational Chemistry
• AI for Science
• Chemical Synthesis

Best for: AI Scientist, Research Scientist, Domain Expert

Related on AIssential

• Apr 29, 2026ScienceEvaluating Claude’s bioinformatics research capabilities with BioMysteryBench — Claude models are rapidly advancing in bioinformatics, matching or exceeding human expert performance on complex, real-world tasks.
• ChemAmp: Amplified Chemistry Tools via Composable Agents — ChemHAS uses hierarchical agent stacking to reduce prediction errors in chemistry tools, outperforming existing LLM and specialist models.
• No hype Claude Opus 4.8 review—my real experience — Claude Opus 4.8 excels in greenfield prototyping but struggles with edge cases and data grounding, often hallucinating.
• Mar 23, 2026ScienceLong-running Claude for scientific computing — Autonomous AI agents can compress months of scientific coding work into days by leveraging structured workflows.
• Claude Opus 4.8 — What’s New? — Claude Opus 4.8 prioritizes "self-honesty" to recognize knowledge limits, reducing confident wrong answers and AI hallucinations.
• On Claude Code with Opus 4.5 — Claude Opus 4.5 excels at complex, multi-file code generation and self-correction in software development tasks.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Anthropic Research.