A July 2026 guide details AI breakthroughs including a model price war that slashed inference costs via Grok 4.5 and GPT-5 iterations, the rise of agentic AI for autonomous multi-step workflows, and a…
A new arXiv paper reveals that PlanFlip attacks achieve a 0.68 attack success rate on GPT-5, with homogeneous multi-agent pipelines using the same LLM backbone for planning and auditing being highly v…
Every major LLM vendor has replaced numeric thinking-token budgets with discrete effort levels, converging on a small enum instead of a number. Anthropic now rejects budget_tokens with a 400 error on …
A new study from arXiv reveals that multi-agent LLM systems are vulnerable to planning-phase prompt injection attacks, with GPT-5 achieving the highest attack success rate (ASR = 0.68) across 3,479 ep…
A study using the ASAP 2.0 dataset found that GPT-5 mini achieved the highest agreement with human ratings for automated essay scoring, while GPT-5 produced the strongest summarization quality, reveal…
A new closed-loop AutoML framework using GPT-5, GPT-4o, and Claude Sonnet 4 as autonomous neural architecture designers achieves mean test accuracies above 93% and a best accuracy of 98.1% for cross-l…
Microsoft Foundry's new Foundry IQ agentic retrieval layer, built on Azure AI Search, treats retrieval as a reasoning task rather than a single keyword or vector lookup. A developer built a production…
Aivolut Books, an AI book-writing app that uses GPT-5, Claude, Gemini, and proprietary tools to turn ideas into structured book drafts and prepare them for Amazon KDP, is offering a lifetime subscript…
A 2017 paper by eight Google researchers, 'Attention Is All You Need', introduced the transformer architecture, which replaced recurrence with attention mechanisms and now underpins nearly all major A…
ChatOn AI Assistant Premium offers a one-year subscription for $29.99 (regularly $39.99), bundling access to GPT, Claude, Gemini, and Perplexity's Sonar models in a single app, compared to ChatGPT Plu…
OpenAI has released a $70 ChatGPT basketball as part of its 'Pause. Play. Prompt.' campaign, alongside a $230 mini keyboard for Codex. The 100% rubber ball is marketed as a physical reminder that crea…
A proof-of-mechanism study of ontology-amplified distillation found that a Qwen3.6-27B student model, fine-tuned on 47 synthetic English-language preference pairs using Foundation AgenticOS ontology, …
Anthropic's Claude Fable 5, the company's most capable publicly available model, achieves top-tier accuracy on eight biomedical benchmarks but refuses to answer between 8.0% and 99.4% of questions dep…
UNIBROWSE, a new unified data pipeline for multimodal browsing, achieves an average accuracy of 54.4 on five benchmarks, outperforming its predecessor Qwen3.5-35B-A3B by 10.5 points and surpassing clo…
OpenAI published a new prompting guide for GPT-5.6 Sol that advises users to stop over-prompting, with internal coding-agent tests showing leaner system prompts improved evaluation scores by roughly 1…
A study in Scientific Reports found that agreeableness is the dominant personality trait promoting cooperation among large language models, with other traits having limited impact. Researchers at Shiz…
A developer discovered that OpenClaw's strongest use case is not AI content generation but content operations pipelines that schedule, adapt per channel, move assets through Git, run ffmpeg, and publi…
A developer building an AI content detector discovered that high accuracy claims are misleading, as even a 0.1% false positive rate can wrongly accuse thousands of real people. The tool's probabilisti…
A developer created an 'Operating Standard' document that encodes a portable doctrine for AI coding agents, enabling cheaper models like Sonnet and Opus to match the conduct of frontier models without…
A Mount Sinai research team led by Mahmud Omar and Eyal Klang found that OpenAI's GPT-5 amplifies sociodemographic biases in clinical decision-making, assigning higher urgency and less advanced testin…