Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
A researcher released DRM Language Emitter, an experimental language model that generates text through learned geometric motion instead of Transformer attention. The project tests whether language can be modeled as contr…
Google engineer Rody Davis demonstrated how AI agents can achieve 100X engineering productivity using Antigravity 2.0, an agent-first platform that unbundles development tools into four pillars: Agent Manager, CLI, SDK, …
Zhipu AI released GLM 5.2, now available via a unified Model API, providing instant access to frontier open-source AI models for building AI apps and agents without deployment. The platform supports multiple models inclu…
BCN, the joint-venture commercial arm of German publishers Hubert Burda Media, Funke and Klambt, launched AI Brand Voice to help brands gain visibility in AI assistants like ChatGPT and Gemini. The product combines AI vi…
A developer tested Agno v2.6.17, a lightweight AI agent framework that rebranded from Phidata in early 2025, using Gemini 2.5 Flash and built-in tools. The framework supports model-agnostic, multimodal, and multi-agent o…
Google Cloud's Open Knowledge Format (OKF) proposes a simple yet powerful approach to enterprise knowledge management by combining markdown content with structured metadata. The format aims to make organizational knowled…
Open-source model GLM-5.2 has surpassed OpenAI's GPT-5.5 in coding benchmarks while costing one-sixth the price, according to its developers. The model features a 1M-token context window and is designed for long-horizon …
GLM 5.2, a new open-weights model, achieved 15% fewer achievements than Gemini 3 Flash in text adventure games, a statistically significant difference. The benchmark, costing $5.1, controlled for game difficulty and foun…
Researchers at an undisclosed institution developed a method to replace attention heads in transformer language models with human-readable Python programs, achieving over 75% similarity in attention patterns on TinyStori…
A developer built AppZ AU Business Advisor, a free tool powered by Gemini 2.0 Flash that generates personalized automation blueprints for Australian small businesses. The tool uses a structured reasoning scaffold with Au…
Pennymac Financial Services is expanding its strategic agreement with Amazon Web Services to deploy enterprise-grade generative AI and commercialize its Plaisse mortgage servicing platform. The deal includes a conversati…
Nexus Labs enabled speculative decoding in vLLM for a fine-tuned 8B model, achieving a 1.9x throughput gain, but discovered that greedy decoding with a draft model is not bit-identical to greedy decoding without one due …
A product manager at a large B2B e-commerce platform applied Andrej Karpathy's LLM Wiki architecture to a professional CRM knowledge base, extending it with source reliability grading and a multi-file schema to handle co…
Chivox released a Model Context Protocol server that turns any LLM into a speech examiner, providing structured scoring for pronunciation, fluency, and rhythm. The tool supports both English and Mandarin, offering granul…
A software engineer argues that developers are overusing large language models for simple data tasks, warning that LLMs are often slower, less reliable, and more expensive than traditional methods like regular expression…
OpenAI added a Scheduled tasks feature to ChatGPT on June 17, 2026, allowing users to set reminders, recurring work, and monitoring tasks, while sunsetting the Pulse proactive updates tool. Pro users have 14 days to migr…
An article on Towards AI explains four advanced ways to use Claude Code beyond basic terminal commands: routines, remote control, scheduled tasks, and channels. These methods help developers maximize productivity by auto…
A developer launched a public World Cup prediction arena pitting 12 AI models against each other, tracking 169 predictions across 21 settled matches. After 21 scoring entries, all models are tied with 3 points each, but …
A team building a production assistant for financial advisors found that prompt editing rarely fixed LLM failures; architectural changes, such as removing abstract routing categories and adding missing tools, were requir…
OpenAI released LifeSciBench on June 17, a benchmark for life-science AI that evaluates applied research tasks like evidence interpretation and experiment design rather than simple recall. The benchmark includes 750 expe…