Blog
The pieces that do not fit in a glossary entry: where a benchmark number came from, what a claim actually shows, and what something costs in production.
Nanbeige4.1-3B: Compact Powerhouse with Strong Reasoning
Nanbeige releases a 4.1.3B parameter model that sets new open-weights benchmarks for efficiency and reasoning on small-scale hardware and mobile devices.
GLM-5: Beyond Vibe Coding to Agentic Engineering
Zhipu AI unveils GLM-5, a state-of-the-art model designed for complex multi-file software engineering and long-horizon autonomous tasks.
Claude Cowork: AI Agent for macOS and Windows
Anthropic's Cowork research preview brings Claude Code's agentic capabilities to everyone, now available on Windows with new global instructions.
NVIDIA PersonaPlex: Controlled Full-Duplex Speech AI
Discover PersonaPlex, NVIDIA's breakthrough in full-duplex speech AI that allows precise control over voice and persona for natural, low-latency interactions.
LingBot-Depth: Precision Spatial Perception for Embodied AI
LingBot-Depth is a high-precision spatial perception model from Robbyant that delivers metrically accurate 3D measurements for robots and autonomous systems.
Youtu-VL: Unified Vision-Language Supervision
Tencent Youtu Lab introduces Youtu-VL, a 4B parameter model that pioneers the 'vision-as-target' paradigm for advanced visual perception.
Tencent HPC-Ops: SOTA Performance for LLM Inference
Tencent releases HPC-Ops, a production-grade high-performance operator library for LLM inference, delivering up to 2.22x speedup on NVIDIA H20 GPUs.
Qwen3-ASR: SOTA Multilingual Speech Recognition and Forced Alignment
Alibaba's Qwen team releases Qwen3-ASR and Qwen3-ForcedAligner, setting new benchmarks in multilingual speech-to-text and precise timestamping.
NVIDIA Earth-2: World’s First Open AI Weather Models
NVIDIA launches Earth-2, a family of open, accelerated models for global weather forecasting, nowcasting, and data assimilation.
HunyuanImage 3.0-Instruct: Tencent’s Massive Native Multimodal Leap
Tencent releases HunyuanImage 3.0-Instruct, the world's largest open-source image MoE model with 80B parameters, unifying understanding and generation.