A new paper shows that simply skewing training toward high-noise timesteps lets diffusion-based robot policies run in a single step, matching 10-step inference without distillation or teacher models.
Datadog AI Research proves time series foundation models scale reliably, releasing five open-weight models from 4M to 2.5B parameters that top every major forecasting benchmark
ProfiliTable is a multi-agent framework that actively explores tabular data before writing code, beating every baseline by 30+ points on a new 18-task benchmark.
A new paper introduces DKPS, a method that reuses cached model responses to predict benchmark scores with far fewer queries than standard evaluation requires.
GPT-Rosalind gets GPT-5.5's agentic coding and deeper domain intelligence for drug discovery, genomics, and wet lab workflows, now open to global enterprise research organizations.
Microsoft's new image model takes #2 on Arena's editing leaderboard and #3 on text-to-image, with both a fidelity tier and a Flash variant priced for production.
Audio-Interaction unifies streaming ASR, voice chat, and proactive sound alerts into one always-on model that decides when to speak from the audio itself
Google's Gemma 4 12B lands in LM Studio with a novel encoder-free architecture that brings near-26B performance, native audio, and multimodal reasoning to any 16GB laptop — for free.
ZipSplat uses k-means clustering to compress scene tokens, achieving state-of-the-art 3D reconstruction with 6x fewer Gaussians and no camera poses required.
Microsoft's first in-house coding model beats Claude Haiku 4.5 on all benchmarks while using up to 60% fewer tokens, now live in GitHub Copilot for VS Code