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 developer argues that looped transformer architectures, such as the described GPT-6 Astra which runs the same blocks through roughly 44 recurrent passes while reusing weights, decouple a model's visible chain of though…
Anthropic released a 154-page threat intelligence report on Thursday detailing how bad actors have attempted to use its Claude AI models for scams, mass surveillance, and biological weapons development, including a May i…
A developer published a tutorial for building a Medical RAG system that parses personal medical PDFs with Unstructured.io, stores embeddings in Pinecone, and cross-references queries against the PubMed API via LangChain …
A developer writing on tamiz.pro argues that agentic frameworks such as the OpenAI Agents SDK expose deep fragility in modern software engineering, coining the term "AI Psychosis" for the chaotic, looping behavior of LLM…
Independent developer croqaz released three new vintage LLM models, a benchmark, and an evaluation pipeline trained only on pre-1900 texts, following the earlier Llama-340M base model that produced Victorian prose but co…
Long-horizon AI agents fail at an exponential rate because a 95% per-step success probability yields only a 35.8% chance of completing 20 turns and 7.7% at 50 steps, and the chance of a second error jumps from 5% to over…
A controlled pilot on a dedicated NVIDIA H200 GPU found ISO-AdamW scored 758 correct answers versus 754 for baseline AdamW on a 1,000-question held-out test, a +0.4% delta with a 95% confidence interval of [−1.8%, +2.7%]…
Anthropic's new documentation for Claude Opus 5 and Fable 5.1 advises users to hand over a complete task specification up front rather than a step-by-step script, warning that skills and prompts written for older models …
InclusionAI is offering free API access to its Ling-3.0-Flash-VL multimodal model and releasing the weights under the MIT license, the company said. The model uses a mixture-of-experts design with 124 billion total param…
OpenBMB released MiniCPM5-2B, a 2.52 billion parameter open-weight language model that posts an average benchmark score of 53.9 across nine capability axes, ahead of Qwen3.5-4B (51.1), granite-4.2-3B (42.7), and LFM2.5-2…
RunPod Serverless lets developers deploy any Hugging Face model as an OpenAI-compatible API endpoint with per-second billing, scale-to-zero, and flash boot cold starts, according to RunPod. The platform runs models on vL…
Together AI expanded its Together Fine-Tuning service to support 17 new open-weight models, including GLM 5.3, DeepSeek-V4-Flash, Kimi K2.7-Code, Qwen 3.5 and Gemma 4, alongside live experiment tracking, finer training c…
Digital Applied published an evidence-classification method on September 12, 2026, for assessing AI benchmark contamination claims, distinguishing training exposure, runtime answer access, and defective grading as separa…
A September 11, 2026 arXiv paper (2609.12303) reports that distilled byte-level transformer models with roughly 1 billion parameters eventually surpass token-based models of the same size, reaching a higher downstream pe…
Researchers Vasileios Baltatzis, Mert Inan, Connor Gillis, Raja Kushalnagar, Lorna Quandt, Leah Findlater, and Colin Lea introduced DiscoSign, a modular Large Language Model-based framework for discourse-aware text to si…
OpenAI published Eric Provencher's blog post "Rethinking skills and prompts for GPT-6 Astra" on September 11, 2026, advising teams to narrow skill descriptions, load supporting guidance only when relevant, revisit compul…
OpenAI introduced ChatGPT for Financial Services on Thursday, an enterprise product built with Morgan Stanley and Evercore as design partners that handles company research, financial data analysis, and pitchbook creation…
Nivedita Mukherjee published an architectural analysis comparing proprietary generative AI abstraction, exemplified by IBM watsonx Code Assistant for Z, against open DevOps orchestration using Zowe CLI, Endevor Web Servi…
Anthropic said on September 10 that it disrupted several allegedly malicious uses of its Claude models over the past eight months, including a Russia-linked cyber espionage campaign and what it called the largest "illici…
GPT-Live-1, a new full-duplex voice API, is now available to developers at $0.05 per minute, allowing the model to listen and speak simultaneously and detect user interruptions in real time. The model scored 80.1% on int…