Tag
Google's Gemini 4 Pro AI model is nearing a preview release with early post-training versions expected to outperform competitors like Astra, with a possible public release in October.
StepFun AI has released a preview version of their Step-5 model in BF16 format on Hugging Face, indicating a promising development in AI model accessibility and performance.
Edge0-35B-A3B-preview is a sparse MoE model that enables efficient AI inference on mobile devices by using streaming expert offloading and quantization, achieving 15 tok/s with under 3 GiB of memory.
Tencent has released the Hy4 preview version, suggesting to maintain a stable release rhythm.
FastVideo releases FastH3 Preview v1, an AI model checkpoint that generates synchronized video and audio from text using four transformer forwards, trained with data-free DMD2 and VSA-H3 at 90% sparsity.
The article explains how the dots3-note Preview model works and provides a link to try it on OpenRouter for free.
Compares DeepSeek V4 Flash preview against the official release, highlighting differences in performance and features.
SupraLabs released SupraWeather-Nano-Preview, a small FT-Transformer model for classifying weather phenomena from tabular meteorological features, trained on synthetic data as an architecture experiment.
Brian Tarricone announces the first preview release of xfwl4, Xfce's Wayland compositor, after six months of work, detailing known missing features and future plans.
The second phase of the LongCat-2.0-Preview closed beta has begun, with limited access slots opening daily at 09:00 Beijing Time.
Alibaba releases Qwen3.6-Max-Preview, an updated version of its Qwen3.6 model series with improved performance and capabilities.