OpenTSLM: The Future of AI Delivered on Time OpenTSLM introduces a new class of foundation AI models designed to interpret and reason about temporal data—time series—as a native modality alongside traditional modalities like text, images, audio, and video. --- Key Highlights What is OpenTSLM? Time-Series Language Models (TSLMs) are multimodal foundation models with built-in support for time series data. They enable AI to understand, explain, and forecast temporal signals, which include heartbeats, stock price ticks, sensor pulses, machine logs, and user clicks. This approach overcomes limitations of current models which lack effective temporal reasoning. Results show order-of-magnitude improvements in temporal reasoning while utilizing smaller, faster model backbones. TSLMs are not an add-on feature but establish time series as a new AI modality. Model Variants OpenTSLM: Lightweight, openly released base models trained on public data; designed to set standards in temporal reasoning and foster a global ecosystem of developers and researchers. Frontier TSLMs: Proprietary advanced models trained on specialized datasets, delivering enterprise-grade performance suited to APIs, fine-tuning, and vertical-specific solutions. Vision Build a temporal interface for AI linking continuous real-world signals to intelligent decision-making and autonomous systems. Enable transformative applications across: Proactive healthcare Adaptive robotics Resilient infrastructure Novel human-AI collaboration forms About the Team Comprised of scientists and engineers from prestigious institutions and companies including ETH, Stanford, Harvard, Cambridge, TUM, CDTM, Google, Meta, AWS. Original authors of the OpenTSLM white paper. --- Resources and Engagement White Paper: Available for download [OpenTSLM-whitepaper.pdf] GitHub Repo: Open source code and models available on GitHub Release Dates: Paper: Sep 30, 2025 Stanford Repo: Oct 1, 2025 Calls to Action Become a Pilot Partner Join the OpenTSLM team Opportunity to Invest Links provided to apply or learn more. --- Interaction Users can describe their work or organization to explore how TSLMs could transform their domain through an interactive text input (currently disabled in the preview). --- Footer © OpenTSLM 2025 --- OpenTSLM bridges AI capabilities with the temporal dimension of the real world to unlock powerful, time-aware intelligence for next-generation applications.