Generative AI news and analysis on Web Pulse: 15564 curated articles tracking the latest Generative AI developments, tools, and research, updated continuously from vetted sources.
Google released Gemini 4 Argon, a new frontier base model that scored 68.9% on the Vals Index, beating Claude Opus 5.5 at 67% and GPT-6 Astra at 63.1%, and 77.9% on DeepSWE v1.1 versus 74.2% and 74.1% for those rivals. A…
The Tokyo District Court ruled Wednesday that an AI-generated voice clone of anime voice actor Kenjiro Tsuda infringed his publicity rights, the first acknowledgment of a right to one's voice in Japan. Presiding Judge Ay…
Google's Gemini 4 Argon raises the model's output ceiling to 1 million tokens, a sixteenfold increase over the previous 64,000-token limit, to sustain long-horizon autonomous tasks. On GraphWalks between 256,000 and 1 mi…
BlackFog launched ADX Vision 2.0, an expansion of its AI security platform that adds prompt protection and governance controls for generative and agentic AI across the enterprise. BlackFog said the release targets risks …
Google's AI Overviews now appear on 93% of branded search queries, according to independent testing by SEO analyst Chris Long, with data from search analytics firm DemandSphere showing branded-query AI Overviews tripled …
Google rolled out skills globally in Gemini chat, letting users save reusable instructions that run by typing a forward slash and the skill name in the prompt bar, and said the feature will reach Google Workspace busines…
Google DeepMind released Gemini 4 Argon, a model it says takes first place on 13 of 19 published benchmarks against GPT-6 Astra and Claude Opus 5.5, with access initially limited to government users and trusted cyber def…
OpenAI released GPT-6.1 Sol (Global) on Amazon Bedrock on 2026-09-29, offering text and image input with a 1,050k-token context window and 128k max output tokens. Amazon lists the model under the identifier global.openai…
Google released Gemini 4 Argon, a frontier model with a 1M-token context window priced at $2/$10 per million input/output tokens, restricting access to "trusted cyber defenders" through its Fairwind Program while running…
Chrome 152 introduces connection allowlists, a security mechanism that lets developers enforce a deny-by-default network sandbox for documents and workers via a Connection-Allowlist HTTP response header. The feature uses…
The Cloud Security Alliance's AI Safety Initiative published a research note on April 19, 2026, finding that 440,445 of 2.23 million code samples generated across 16 code-generating models in Python and JavaScript — 19.7…
Pixability launched Pixie, an agentic tool that screens YouTube content for AI slop, assessing some 8 billion videos across 40 million channels using a combination of human verification and AI, chief product officer Jack…
Google announced Gemini 4 Argon, a frontier AI model it says can locate, validate, and patch critical software vulnerabilities without human help, and is rolling it out to trusted cyber defenders through its Fairwind Pro…
EXPLORENET, a learned adaptive exploration policy for group-relative diffusion RL, improves held-out GenEval2 by 14% over Flow-GRPO on Stable Diffusion 3.5 Medium and reaches a 67.2% human preference win-rate, according …
A study of 20,462 student turns from 1,260 authentic sessions with a guided LLM chemistry tutor found that each additional impasse turn lowered the odds of next-turn recovery by 12.7% (AOR = 0.873, p < .001), according t…
Researchers introduced modal kinetic typography, a method that animates a vector glyph by building motion from the glyph's natural vibration modes, computed via a finite-element eigenproblem assembled from the vector out…
Google unveiled Argon, a new flagship AI model for its upcoming Gemini 4 generation that the company says is its highest-performing model to date and larger than its previous high-end Pro models, with availability initia…
Developer Rehan launched Universal Modder, an open-source toolkit that lets AI coding agents mod almost any PC game through a shared knowledge base, Agent Skills, a fal MCP server and a `um` CLI. The project supports Cla…
A developer published a reproducible diagnostic showing that lowering the volume of an already-clipped AI-generated audio export cannot reconstruct the flattened waveform, using a standard-library Python script and FFmpe…
A software engineer and AI master's student used Claude Opus 5.5 in Cowork to produce a 45-second career-switch video entirely from code, iterating through nine versions and 666 frames rendered headless with Playwright a…