Tag
SandboxAQ has integrated its large quantitative models (LQMs) for drug discovery and materials science into Anthropic's Claude, enabling researchers to use these powerful tools through a natural language interface without specialized computing infrastructure.
This paper introduces ChemCost, a benchmark for evaluating how well LLM agents can estimate chemical procurement costs by grounding identities, retrieving quotes, and handling noise. It reveals that current agents struggle with robustness and precise arithmetic reasoning in scientific workflows.
OpenAI releases GPT-Rosalind, a specialized life sciences model optimized for protein reasoning, chemical analysis, genomics, and scientific workflows.
AIPOCH Open Science is an open-source, local-first AI research workbench for reproducible science featuring scientific AI agents, Python and R execution, and cross-platform support. The content announces the v0.25.1 maintenance release, which includes fixes for protected notebook kernels on Windows and atomic file exports with time-zone-safe timestamps.