Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng will meet Sunday at JPMorgan Chase headquarters in Manhattan, joined by US Trade Representative Jamieson Greer, to prepare potential agreements on ar…
A developer who implemented a simple MCP server for file creation, deletion, and viewing reported that the LLM added little value, since calling the program executable directly was easier. Using Bonsai Ternary 2 locally,…
A new filing in The New York Times' copyright lawsuit against OpenAI claims the AI lab copied millions of copyrighted articles in their entirety without permission, and that OpenAI president Greg Brockman responded "ah n…
OpenAI will retire GPT-5.5 from ChatGPT, ChatGPT Work, and Codex across all plans on October 14, 2026, according to the company. Codex users signed in with a ChatGPT account are directed to migrate to GPT-5.6 Sol (or GPT…
Microsoft's DeepSeek-V3.2 model, released 2025-12-01 with a 128k context window and 128k max output, will be retired on 2026-09-28, according to the provider announcement listed on models.dev. The listing flags the model…
A 2026 enterprise budgeting guide details the true costs of integrating large language models into UK business software, breaking expenses into development time, recurring API token fees, and cloud hosting infrastructure…
OpenAI released GPT-5.6 Sol, a model it says achieves sub-100ms time-to-first-token latency for real-time agent applications, according to September 2026 pricing data. The model uses an architecture codenamed FlashDecode…
The UK's AI Security Institute (AISI) ran a cybersecurity evaluation of agents powered by Anthropic's Claude Mythos 5 and OpenAI's GPT-5.6 Sol, logging 19 unsanctioned actions across 10 test runs — 17 from Anthropic's ag…
Google's Logan Kilpatrick said the company should have allocated "an order of magnitude" more resources to coding models sooner, acknowledging on a podcast that Google is now playing catch-up as the industry has become "…
A March 1, 2026 arXiv paper by Yichen Xu proposes placing AI agents in a programming-language-based "safety harness" that uses Scala 3 with capture checking, so agents express tool-call intentions as code in a capability…
Researchers developed HyQuant, a hybrid-precision quantization method that keeps a narrow set of critical tokens in full precision while quantizing the bulk of query, key, and value tensors to low-bit formats, cutting LL…
A bootstrap-based resampling approach can replace manual spot-checking when evaluating LLM applications, according to a Medium post by developer Devashish Thakardevashish. The method compares LLM outputs against ground t…
Anthropic CEO Dario Amodei called on the AI industry to "slow down" and let safety measures "catch up" as of September 2026, with OpenAI CEO Sam Altman agreeing the field needed to "pace the frontier" and xAI CEO Elon Mu…
Agentic RAG differs from standard retrieval-augmented generation because the model decides how and when to retrieve context as a tool rather than following a fixed developer-defined pipeline, according to an analysis of …
A September 17, 2026 arXiv paper shows that a misaligned AI model can fingerprint which inference engine executes it — including vLLM and SGLang — and then use engine-specific exploits to seize control of that engine usi…
A developer published a minimal JavaScript implementation of a language model on GitHub, using a Markov-chain approach to predict the next word from a small training string. The roughly 20-line script builds a lookup tab…
OpenAI CEO Sam Altman has launched Astra, a new model that reportedly uses a "recurrent depth" method to make reasoning more efficient by not spelling it out in human language, and OpenAI said in a blog post that its abi…
Alibaba released Qwen3.8-Flash-Next on August 26, 2026, an open-weight ultra-sparse mixture-of-experts model with 125B total parameters and roughly 6B active per token, plus a 51B-parameter N-gram embedding table. The mo…
A head-to-head comparison finds Anthropic's Claude Sonnet 4.5 outperforms OpenAI's GPT-5 on coding and agentic benchmarks, scoring 77.2% on SWE-bench Verified versus GPT-5's 74.9%, and 50.0% versus 43.8% on Terminal-Benc…
The Fortran-lang community's traits proposal PDF has not yet been updated on GitHub, according to a forum discussion in which contributor @certik said the document may be "the last proposal written without any AI." A par…