Alibaba's Qwen team released Qwen3, a new large language model available in sizes including a 27B version, with official BF16 and FP8 weights on Hugging Face and community GGUF and MLX quantizations f…
OpenAI's gpt-oss-120b model, with open weights, requires 72 KiB of KV cache per token in 16-bit precision, calculated from its config.json with 36 layers, 8 key-value heads, and a head dimension of 64…
Alibaba Group Holding's open-weight Qwen AI models have surpassed 3 billion global downloads in the past six months, making them the world's most-downloaded AI models, according to a Hugging Face repo…
A daily AI news digest published on Aug 16 lists 30 stories, including reports that OpenAI is losing key personnel ahead of a potential IPO, Qwen 3.8 27B outperforms the larger Qwen 3.7 Plus in coding…
Qwen's Qwen3.8-27B model has been crowned the best laptop-runnable model by community testers, though users report it overthinks and generates 5x more tokens, while GLM-5.3 from Z.AI is praised for it…
A developer has created a benchmark harness for testing open-weight models against real product tasks before routing production traffic, addressing the hidden costs of cheaper models that fail in prac…
Klein Pass, a $9.99-per-month subscription from the team behind the Klein VS Code coding extension, bundles discounted API access to open-weight AI models including Kimi K3, Kimi K2 series, DeepSeek V…
Kortexio released ContextMemory v0.1.0-beta, an agent memory system designed to be opened like a wiki rather than a vector black box or classic RAG inject. The release adds Cursor-style HTTP, vision, …
HuggingFace's State of Open Models report for summer 2026 reveals that Chinese labs now dominate frontier-scale open models, with the largest reaching 2.78 trillion parameters, while US labs peaked at…
Large language models (LLMs) function as massive pattern libraries for mathematical proofs, excelling on familiar problems but failing on novel twists, according to a technical analysis. The article a…
Apple trained its own AI model for China with support from Alibaba, according to Reuters, and the Cyberspace Administration of China registered Apple's generative AI service last month, clearing a key…
Alibaba's Qwen team released Qwen3.8-27B, an open-weight multimodal model with 27 billion parameters and a native 262,144-token context window, available under Apache 2.0 on Hugging Face and ModelScop…
Alibaba's Qwen 3.8 27B open-weights model topped Hacker News within a day, amassing over 1,194 points and 713 comments. The dense 27-billion-parameter model, Apache 2.0 licensed, can be run locally on…
A technical tutorial demonstrates an end-to-end supervised fine-tuning pipeline for the XYZ-Aquila-SFT dataset using Hugging Face Transformers, PyTorch, and PEFT, culminating in LoRA fine-tuning of Qw…
Alibaba's Qwen family of AI models has surpassed 3 billion global downloads, accounting for more than 50% of all open-source model downloads worldwide, outpacing Meta's Llama and DeepSeek. The milesto…
Alibaba Group Holding's open-weight Qwen models surpassed 3 billion global downloads in the past six months, making it the world's most-downloaded AI model family, according to a Hugging Face Inc. rep…
Klorn, an email classification service, suffered a cost cap bug that over-billed by up to 100x due to substring matching in its pricing table. The developer had previously fixed the table to over-esti…
Qwen released Qwen 3.8 27B, a dense model that runs at approximately 138 tokens/second on an RTX 5090 with the ninfer inference engine, and Z.AI released GLM-5.3, a coding-focused model that achieves …
European enterprises are actively testing Chinese AI models on local servers, challenging Brussels' technological sovereignty drive, according to Volker Pfirsching, a Munich-based partner at Arthur D.…
Qwen 3.8 27B, a 27-billion-parameter language model, has been released on the Hugging Face platform, offering improved performance for natural language processing tasks such as text generation, transl…