Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Apple shipped iOS 27 on Monday, September 14, 2026, with Siri now running on Google's Gemini 2.5 Pro under a partnership reportedly worth roughly $1 billion a year to Google, while the European Union is excluded from the…
Google announced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking on September 15, 2026, positioning the models for production voice agents with native speech-to-speech processing, asynchronous function calling, nea…
Saudi Arabia's state-backed Humain launched the Arabic-language AI model Humain-m3, built on an open-source system from Chinese startup MiniMax, even as the kingdom spends billions on AI infrastructure with U.S. companie…
A developer has proposed TokenPot, a community-pooled model for purchasing LLM inference in which 1,000 developers each contribute $5 per month to collectively buy compute at volume pricing and distribute it via an OpenA…
Fatih Kadir Akın launched Awesome ChatGPT Prompts in December 2022, and the project has grown into prompts.chat, an open-source prompt repository and MCP server ecosystem that now holds over 170,000 GitHub stars. The rep…
Mozilla announced on September 16 that Mistral Small 4 will power Firefox Smart Window, its optional AI browsing assistant, for beta users in the US and Canada, with France becoming the first market to receive official F…
Apodex released Apodex 1.1 on August 24, a model family whose Deep Discover mode assembles an asynchronous "Agent Team" that splits tasks, runs analysis code on files, and returns traceable tables, charts, and scripts. O…
Google released Gemini 3.8 Live, an "Extended Thinking" version of its conversational AI that can run background tasks such as drafting emails, retrieving data, and scheduling meetings during a conversation, according to…
LatticeFlow AI released the first independent framework for objectively measuring political bias in large language models, producing a reproducible bias score with cryptographic verification of the tested weights. In its…
An analysis of 4,749 public MCP servers found that the most common reason language models fail to call a registered tool is a description that lacks any distinguishing term from its siblings, affecting 17.7% of public to…
A developer's cost experiment comparing an LLM-powered support agent against deterministic code found that a refund-eligibility check run 50,000 times a day costs roughly $14,600 to $73,000 per year via Claude models ver…
GAVEL, a judge protocol described in a paper on arXiv, reduced evaluated-timeline discrepancies from 7.63 to 0.85 per clinical case report after LLM-guided merging, with merged timelines preferred in 77.0% of comparisons…
OpenAI, Anthropic, and Google DeepMind have been coordinating on AI safety for weeks, discussing third-party evaluators and a possible industry standards body for frontier models, according to TechCrunch. The talks signa…
A developer traced the network calls of an agentic coding assistant and found that a simple two-file PyQt edit consumed 3 LLM calls and roughly 760 KB of exchanged JSON, versus a single call and about 100 KB with Aider. …
The Beijing Economic and Technological Development Zone (Beijing E-Town) has launched a "token loan" program in which six Chinese banks — Agricultural Bank of China, China CITIC Bank, Industrial Bank, Bank of Beijing, Ch…
Mozilla's Firefox Smart Window AI browsing assistant is now powered by Mistral models, the two companies announced, with the beta rolling out to users in France and North America and the United Kingdom and Germany expect…
Uncensored local versions of Google's Gemini and Alibaba's Qwen models answer 95% to 100% of unsafe questions, compared with 1% to 5% for the base models, according to an experimenter who replicated the findings and ran …
A developer built model-radar, an open-source tool that automatically discovers free LLM endpoints and adopts qualifying models into an agent stack, but only after they pass a fixed five-stage deterministic gate. The gat…
Apify, Firecrawl, and Crawl4AI lead a 2026 roundup of eight AI data collection tools for building model-ready datasets, with Apify priced from $19/month, Firecrawl from $16/month, and Crawl4AI free and open-source. The t…
Google's Gemini for Go Developers series published Part 3, a tutorial on building AI agents in Go using the Go GenAI SDK, Genkit, and Google's Agent Development Kit (ADK). The article defines an agent as a large language…