🎉 2026-02-14 · v0.1.3 Released. The v0.1.3 release introduces full support for the latest GLM-5 model, achieving up to 500 tokens/s on GLM-5-FP8 and up to 600 tokens/s on DeepSeek-V3.2. TileRT is a ...
ThunderKittens is a framework to make it easy to write fast deep learning kernels in CUDA. It is built around three key principles: ThunderKittens is built from the hardware up; we do what the silicon ...
Copilot in Excel introduces new finance-focused capabilities that help teams work to their standards, use trusted financial data, and stay in control of every change. GPT‑5.5 Instant is now available ...