Aleph Alpha released Kolibri on October 3, 2026, a 78.1-billion-parameter open-weight Mixture-of-Experts model with a context window of up to 1,048,576 tokens, trained on 768 NVIDIA B200 GPUs in Germa…
Tavus launched Griffin-Lite on October 1, 2026, and 48% of 54 participants in a one-minute video call believed they were speaking to a real person, up from a previous best of 2.4% for any AI video sys…
BoxLite ranked first of ten projects in HVTracker's Sandboxes & Runtimes category on October 3, 2026, with an HVTrust score of 85.6 — the only A grade in the category, ahead of E2B (78.2) and NVIDIA's…
A scheduled inspection of a raised-floor plenum beneath a 2-megawatt AI cluster of 128 liquid-cooled racks, each holding eight NVIDIA H100 GPUs, uncovered zinc whiskers growing on galvanized floor sup…
Kazakh Prime Minister Olzhas Bektenov told the AI & Digital Bridge 2026 forum in Astana that Kazakhstan has the legal framework in place for large-scale artificial intelligence adoption, citing the co…
Prime Intellect launched Prime Inference, a serverless and reserved-capacity serving platform for frontier open-source models that processed nearly a trillion tokens per day internally before public r…
A new arXiv paper (2610.00049v1) reports that replacing native sigmoid, tanh, and SiLU activations with degree-3 or degree-4 bfloat16 polynomial programs improved complete training-step throughput on …
Tavus announced Griffin, which it calls the first model to pass the video Turing test, with 48% of people who talked to it live believing it was a real human, versus a pass rate below 3% for previous …
Developer Ben Houston released three-dlss-nr, a Three.js port of maanHimself's OpenDLSS-NR that runs the 71-block DLSS 5 neural rendering network as TSL compute kernels inside a WebGPURenderer with no…
The OpenClaw Foundation launched OpenClaw Enterprise (OCE) on September 29, a free, MIT-licensed, self-hostable control plane for persistent AI agents backed by OpenAI, Red Hat, and NVIDIA. OpenAI ori…
NaiveAI released Naive-N0.5-Flash, an MIT-licensed open-weight mixture-of-experts coding model with 309 billion total parameters that activates 15.5 billion per token and ships with a native 1-million…
Janus embeds llama.cpp directly inside a lightweight Go binary with a Vulkan acceleration bridge, eliminating the Python runtime and Docker virtualization layers used by typical local LLM stacks. The …
Fixstars CTO Takuro Iizuka reported at Autonomous Vehicles & AI USA 2026 that an AI-agent-assisted rewrite of an autonomous driving training codebase cut training time 1.7x, with the change spanning r…
MSI announced a 64 GB configuration of its EdgeXpert desktop AI supercomputer, built on the NVIDIA DGX Spark platform, complementing the existing 128 GB tier. The 64 GB model runs up to 100B parameter…
CoreWeave put NVIDIA Vera Rubin NVL72 into production at its Fully Connected 2026 conference in San Francisco, with Cognition as the first customer running live workloads; Cognition engineers report u…
A Proxmox VE 9 user running kernel 7.0 documented a working procedure for passing NVIDIA Tesla A2 GPUs through to privileged LXC containers with CUDA support on Debian 13 Trixie, pinning the NVIDIA dr…
NVIDIA will release a 64GB unified-memory configuration of its DGX Spark personal AI supercomputer this month through manufacturer partners Acer, ASUS, Dell, Gigabyte, HP and MSI, retaining the GB10 G…
NVIDIA has adopted the "Ozaki Scheme II" computational method, developed by Professor Katsuhiro Ozaki and his team at Shibaura Institute of Technology, into its CUDA toolkit 13.4, according to EE Time…
A procedural game called Hops Run topped its leaderboard with a Qwen3_4B/1 agent from SemIf reaching 6330 m across 814 runs, ahead of NVIDIA's Kumo_Tabular/1 small agent at 5994 m, both on game versio…
Clustering two NVIDIA DGX Sparks over a 200 GB RoCE RDMA cable and running tensor-parallel inference with vLLM lets the pair serve models in the 150 to 160 GB range, such as DeepSeek V4 Flash, that do…