↗
AMD Adds Day 0 Support for Qwen 3.8 27B on Ryzen AI Max
AMD ships Day 0 support for Qwen 3.8 27B, letting Ryzen AI Max+ PCs and a Radeon AI PRO R9700 run the dense model locally at up to 51.8 tokens per second.
CH CHANNEL ARCHIVE
Tag: AI
Artificial intelligence models, systems, and products.
01 INCOMING
ARCHIVE
↗
↗
Anthropic Details How Claude’s Text Watermarking Will Work
Anthropic explains how Claude's text watermark works, why the EU AI Act drove it, and what the technique can and cannot prove about generated text.
↗
Z.ai Ships GLM-5.3, Its Strongest Open-Weight Coding Model
Z.ai's GLM-5.3 keeps the GLM-5.2 base and adds post-training only, posting top open-source coding scores and a sharp jump in cyber-security benchmarks.
↗
Google Launches Gemini 3.7 Flash With Half-Price API Tier
Google's Gemini 3.7 Flash lands three weeks after 3.6 Flash with stronger coding benchmarks and an introductory API rate priced at half the previous level.
↗
Google’s Tensor G6 Brings Faster On-Device AI to Pixel 11
Google's Tensor G6 chip powers the Pixel 11 series with 50 percent more TPU compute and up to 3.5x faster on-device AI, plus post-quantum secure boot.
↗
Lam Research Plans $3 Billion Chip R&D Lab Expansion
Lam Research will invest more than $3 billion over five years to expand its global R&D lab network and lift experiment capacity by more than 50 percent.
↗
IBM and OpenAI Form Enterprise AI Partnership With GPT-5.6
IBM will embed OpenAI's GPT-5.6, Codex, and ChatGPT Work into IBM Consulting Advantage and launch a dedicated OpenAI Practice.
↗
DeepSeek Ships Official V4 Pro With DSpark Decoding
DeepSeek ships the official V4 Pro release (0813) with MIT-licensed weights, DSpark speculative decoding, and large agentic benchmark gains.
↗
Nvidia Releases Nemotron 3.5 Lightning for Agentic AI
Nvidia's 30B-parameter Nemotron 3.5 Lightning open model targets long-running agentic workloads, joined by the NeMo Switchyard routing library.
↗
CME Group Sets October 5 Launch for GPU Compute Futures
CME Group will launch the first regulated compute futures on October 5, tracking Nvidia H100 and B200 rental prices on NYMEX.