Google Cloud Pairs Quantitative Models With Gemini for Science Push

Google Cloud will begin selling specialist AI models built for scientific computation, adding SandboxAQ's large quantitative models to its marketplace alongside Gemini. The announcement, made June 29, 2026, targets research workflows in drug discovery, materials science, and semiconductor manufacturing where general-purpose large language models routinely produce inaccurate numbers.

Large language models excel at generating fluent text but remain unreliable with arithmetic and structured scientific data. SandboxAQ's large quantitative models, by contrast, are trained on numerical data and scientific equations rather than prose, a design choice meant to improve accuracy in chemistry, biology, and physics problems where the correct output is a number or molecular structure, not a paragraph.

Gemini for Science Bundle

Google paired the marketplace listing with Gemini for Science, a curated set of tools and experimental agents aimed at the research workflow. The bundle includes the AI co-scientist, the AlphaEvolve coding agent, an empirical research assistant, and NotebookLM. Google framed the offering as a way to accelerate routine, laborious steps of the scientific method rather than to replace the scientist.

• Researchers can combine Gemini's reasoning and interface capabilities with SandboxAQ models for the underlying scientific computation.
• DeepMind's protein-structure work has already influenced drug development, and a separate effort produced an AI that found more new materials in a year than science had catalogued in its entire history.
• Google said the capabilities are already in use by partners in private preview for real-world R&D, though it has been sparing with specifics on which organisations and what results.
• DeepMind's drug-discovery spinoff Isomorphic Labs is moving toward trials, and rivals across the industry are racing to convert algorithmic promise into treatments and materials that work outside a benchmark.

Strategic Race for Science AI

The commercial logic is straightforward. Google is competing with other hyperscalers to be the default platform for enterprise AI, and scientific and industrial R&D represents a high-value segment that general chatbots do not serve well. Selling specialist models through the marketplace, the same channel through which it already offers a wide catalogue of third-party systems, lets Google capture that demand without having to build every domain model itself. Whether the combination produces discoveries or simply faster spreadsheets is the question the private previews are meant to answer. The marketplace listing is the substantive change: a category of AI that was largely confined to specialist labs becomes something a research team can rent.