SpaceXAI released Grok 4.5, a general-purpose AI model trained alongside Cursor for coding, agentic tasks, and knowledge work, priced at $2 per million input tokens. The model claims superior token ef…
SpaceXAI and Cursor launched Grok 4.5, a mixture-of-experts model co-trained on developer-agent interaction data, prioritizing token efficiency and aggressive pricing over benchmark dominance. The mod…
XAI released Grok 4.5, a large language model trained on tens of thousands of Nvidia GB300 GPUs. Despite trailing Fable 5 and GPT-5.5 in coding benchmarks, Grok 4.5 costs $2 per million input tokens, …
SpaceXAI released Grok 4.5, a new AI model trained alongside coding tool Cursor, aimed at coding and technical tasks. Cursor co-founder Michael Truell called it an "Opus-class model" that is fast and …
Supermicro showcased its Super AI Station, a DGX Station system powered by NVIDIA's GB300 Grace Blackwell chip, at Computex 2026. The liquid-cooled desktop features 252GB of GPU memory, 496GB of CPU m…
HIVE Digital Technologies signed a non-binding letter of intent with a sovereign Swedish technology company to lease 25 MW of its Boden, Sweden facility for high-performance computing colocation over …
SGLang achieved a 5x throughput improvement for serving DeepSeek-V4 on NVIDIA GB300 hardware since its Day-0 launch, reaching ~11,200 tok/s/GPU at 50 tok/s/user through kernel and runtime optimization…
A developer used Gemini Flash Lite to transcribe a video explaining High Bandwidth Memory (HBM) technology. The video details how HBM achieves high bandwidth through an extremely wide bus width, contr…
Startup Tensordyne claims its new Napier AI chip, using logarithmic math to replace multipliers with adders, can run large language models four times faster using one-fifth the power of a comparable N…
Artificial Analysis released initial results for its AA-AgentPerf benchmark, showing NVIDIA's Blackwell systems outperforming AMD's Instinct MI355X GPUs on power-efficient agentic inference using Deep…
SpaceX has nearly completed version 1.0 of a proprietary AI training stack written in C, designed to exact-map to 220,000 GB300 processors with 800G network interface cards. The system, which leverage…
Mistral has launched Mistral Cloud, a private, integrated compute stack featuring GB300 GPUs, 1:1 InfiniBand XDR fabric, and SLURM plus Kubernetes orchestration. The platform offers tiered deployment …
PyTorch 2.11 now enables direct installation of CUDA-enabled PyTorch wheels on aarch64 Linux from PyPI, eliminating the need for custom package indexes and workarounds that previously complicated depl…