Explore the complete news stream by source, topic and date.
llama.cpp v.b11861 fixes the weight distortion issue in the minimax-m3 model that occurred under NVFP4/MXFP8 quantization due to the scale tensor not being injected into the computation graph, thereby preventing “token soup” and residual vanishing; it also provides multi-platform binary builds covering macOS, Linux, Windows, Android, iOS, and openEuler.
Introducing Kimi K3.1 Pro Kimi K3.1 Pro landed on the Kimi API today - it's not yet open weights, but Moonshot promise "Full model weights will be released by Nov 4, 2026". They're also offering a "Highspeed" version that's 2-3x faster, at twice the price (except for input token cache hits, which are priced the same). The US dollar pricing is $2.00/million tokens for input and $10/million for output - the same as GPT-6.1 Sol and half the price of Claude Opus 5.5. Moonshot's own benchmarks show Kimi as approximately level with Opus 5.5 and GPT-6 Astra - beating them on a few benchmarks, coming close on others. Artificial Analysis scored Kimi at 55 - ahead of Claude Fable 5.1, GPT-6 Astra, and Gemini 4 Argon (all at 53) and behind only Claude Opus 5.5 (58). This is the first time a (soon-to-
On October 22, 2026, the U.S. Federal Communications Commission (FCC) granted Intel a one-year exemption, allowing it to continue using certain foreign‑made automated guided vehicles (AGVs) at its U.S. semiconductor fabs, as the original ban’s vague wording could inadvertently affect critical semiconductor manufacturing equipment.
Kazakhstan’s AI ecosystem is transitioning from the planning phase to large-scale deployment. Firebird has obtained U.S. export licenses to build an AI factory in Kazakhstan, with the first phase deploying 45,000 NVIDIA Grace Blackwell GPUs. Kazakhtelecom has already brought online more than 1,000 H200/B300 GPUs and plans to add over 2,000 Vera Rubin and other GPUs by 2025. Multiple stakeholders are collaborating to advance computing power supply, developer access, and the local implementation of AI applications.
AWS has announced that SageMaker AI’s serverless model customization feature now includes “Training Sessions,” which are fully available starting today. Users can independently define post-training logic—such as custom loss functions, reward signals, or multi‑round reinforcement learning—while the platform automatically handles GPU allocation, distributed execution, fault tolerance, and checkpoint storage.
Amazon SageMaker has released the Training Sessions API, enabling users to implement custom large-model training workflows—such as SFT, RLHF, GRPO, and KTO—using pure Python scripts. GPU resource scheduling, fault tolerance, distributed execution, and checkpoint management are all handled by the service.
Elastic GPUs, sandbox growth, and the economics of the agent cloud
Elastic GPUs, sandbox growth, and the economics of the agent cloud
European retailers are leveraging NVIDIA-accelerated open large models and physics‑AI technologies to achieve large‑scale deployment across scenarios such as digital product discovery, intelligent warehouse simulation, virtual try‑on, and accelerated R&D. Companies including Shopify, Zooplus, L’Oréal, and the Otto Group have used tools like the Nemotron series of models, Omniverse digital twins, and ALCHEMI molecular simulations to significantly reduce inference costs, improve search accuracy, shorten R&D cycles, and validate improvements in conversion rates and sales growth.
Jas Kaira, head of Blackstone’s AI business, stated at the TechCrunch Disrupt conference that market panic over massive debt in the AI infrastructure sector has been overly amplified. His key argument is that this debt is largely backed by long-term contracts with high‑credit‑rating clients—such as Microsoft and Meta—and that the underlying assets (e.g., GPUs) possess strong collateral value, meaning the risk profile differs fundamentally from that of the 2008 banking system.
A U.S. federal judge ruled that the Trump administration lacked the authority to instruct the Environmental Protection Agency (EPA) to unilaterally dismiss an air pollution lawsuit against SpaceX AI over its operation of an unauthorized gas turbine in Memphis; the lawsuit filed by the NAACP will proceed.
On October 26, 2026 (UTC), NVIDIA announced that the cuDSS library has entered General Availability (GA). This CUDA-X library is specifically designed to efficiently solve large-scale sparse linear systems on GPUs, significantly boosting simulation throughput and iterative efficiency across fields such as chip design, aerospace, and industrial systems.
Dharma-AI has introduced the VocabX method, which reconfigures the vocabulary of language models for specific tasks—such as automated scoring of Portuguese essays—by optimizing for cost rather than word frequency. Without modifying the model architecture, parameters, or training pipeline, empirical results show a 77% reduction in inference costs, while the evaluation score (bench score) improved from 0.703 to 0.810.
This article is the first in a series of technical blog posts jointly published by Hugging Face and Amazon, systematically introducing how to use TorchNeuron on AWS Trainium AI accelerator chips for PyTorch model inference performance analysis. It covers the principles of Neuron backend integration, methods for configuring torch.profiler, the use of NeuronConfig and NeuronProfiler tools, and visualized tracing with Neuron Explorer. Additionally, using RMSNorm as an example, it compares the device‑time differences among three implementations: plain PyTorch, the built-in torch.rms_norm fusion, and torch.compile’s automatic fusion.
i14 has launched the Laplace 3B model with 3.028 billion parameters, which replaces Softmax attention with Gated Linear Attention (GLA). During inference, its per-sequence recursive state remains constant at 29.5 MB, independent of context length. The model was trained and validated on a single AWS trn2.48xlarge instance (equipped with 16 Trainium 2 chips), achieving a training throughput of 120,000 tokens per second and an inference throughput of 11,544 output tokens per second (with 1,280 concurrent sequences). All performance metrics are based on internal measurements conducted by i14 and the AWS GenAIIC team.
SF Compute was founded by Evan Conrad and is dedicated to building a neutral, programmable GPU computing‑power trading infrastructure that supports physical delivery. Its model rejects the abstraction of FLOPS as a commodity, emphasizing hardware heterogeneity, actual delivery, and real‑time constraints—such as interrupt tolerance, InfiniBand topology, and firmware versions. The platform has already facilitated tens of billions of dollars in transaction volume and introduced low‑level capabilities like dynamic InfiniBand subnet partitioning.
SF Compute was founded by Evan Conrad and focuses on enabling programmable, physically settled trading of GPU computing power. It does not operate its own data centers; instead, it leverages full-stack technology to collaborate with third-party facilities in co-building a “sovereign compute cloud,” emphasizing market neutrality and infrastructure controllability. The company has already facilitated hundreds of millions of dollars in transaction volume and implemented key capabilities such as dynamic InfiniBand partitioning.
To meet the surging demand for high-bandwidth memory chips driven by the explosive growth of generative AI, Samsung is building a semiconductor manufacturing cluster in Pyeongtaek, South Korea, spanning an area equivalent to New York’s Central Park. To date, three fabs have been commissioned, with a fourth set to begin production soon, and construction on a fifth—halted in 2024 due to declining demand—has now resumed. The site showcases an intense construction pace and a workforce ecosystem adhering to stringent safety standards.
Hugging Face has released TorchTPU, providing native TPU backend support for PyTorch, enabling libraries such as Transformers and Diffusers to run directly on TPUs without relying on torch_xla or migrating to JAX.
Hugging Face has officially released AnyLanguageModel 1.0, an open-source Swift framework fully compatible with the Apple Foundation Models API. It supports seamless switching between local models (MLX, llama.cpp, Core ML) and hosted models (OpenAI, Anthropic, Gemini, Hugging Face Inference Providers, etc.), and introduces new features such as image input, tool invocation, guided generation, and automatic adaptation of inference chunks.
