MiniMax Group Inc | Reuters
reuters.comChinese AI startup MiniMax posted a nearly four-fold jump in first-half revenue as demand for its low-cost models accelerated. The company released new video-generation and large language models, seeking HK$16 billion in fresh capital via share sale to fund growth in its AI business.
China's Moonshot unveils world's largest open AI model, closing in on US rivals | Reuters
reuters.comChinese AI startup Moonshot unveiled Kimi K3, a 2.8 trillion-parameter open-weight model that delivers performance approaching US rival Anthropic's Fable model. The launch highlights China's rapid advancement in the AI sector with strong results on third-party benchmarks and architectural upgrades for computing efficiency.
Perplexity Launches Hybrid Compute on Mac With Local Privacy Gate – Unite.AI
unite.aiPerplexity introduced hybrid compute for the Perplexity Mac app, dividing each Computer task between frontier models in the cloud and a local model running on the Mac. Available to Pro, Max, and Enterprise subscribers with three local models: Gemma 4 E4B, Qwen3.6 35B-A3B, and a Perplexity post-trained model.
Nvidia Eyes New Perplexity Investment at More Than $30 Billion Valuation | citybiz
citybiz.coNvidia is in discussions to invest more than $30 billion in Perplexity, valuing the AI startup at over $30B as its annualized revenue grows from under $250M to over $750M. The investment would mark a 50% increase from previous valuation and illustrates Nvidia's strategy of investing across the AI ecosystem.
NVIDIA Vera Rubin Ramps Into Full Production to Power Agentic AI Factories Worldwide
investor.nvidia.comNVIDIA announced the Vera Rubin platform is ramping into full production to power agentic AI factories worldwide, with Taiwan's top server makers and global supply chain leaders manufacturing systems at scale.
SpaceX shares sink after first earnings report reveals huge AI spending plans - BBC News
bbc.comBBC article covering SpaceX's Q1 2026 earnings showing quarterly revenue of $7.8bn with capital expenditures hitting $18.4 billion, where bulk ($15.8bn) was for AI infrastructure and compute deployment including Colossus II data center build-out. Stock tumbled after revealing huge AI spending plans to investors.
Gemini Spark AI Agent Announced at Google I/O 2026 - AP News
apnews.comGoogle I/O 2026 introduced Gemini Spark, powered by Gemini 3.5, as a cloud-based AI agent that completes mundane tasks like sorting emails and creating summary documents from meeting notes. The proactive nature of this agentic AI differentiates it from traditional chatbots while being designed to ask permission before performing high-stakes actions like sending emails or making purchases.
Why Google's Gemini Spark AI agent could be a game changer - CBS News
cbsnews.comGoogle announced Gemini Spark, an always-on cloud-based AI agent that can complete routine tasks like sorting emails and creating documents with meeting takeaways. Unlike other agents requiring local hardware, Spark runs in the cloud continuously even when users shut down their devices. The tool represents a paradigm shift from chat interfaces to autonomous task completion for consumers entering agentic AI territory.
ToolGrad: Efficient tool-use dataset generation with textual gradients
research.googleGoogle Research introduces ToolGrad for efficient tool-use dataset generation using textual "gradients" to achieve high annotation pass rates. This advancement in generative AI methodology addresses key challenges in building robust multi-modal and agent-based systems.
Google's July 2026 AI Updates - Faster Gemini Models and Advanced Robotics
blog.googleJuly 2026 updates introduce faster Gemini models, advanced robotics capabilities, and new creative tools for video and music production. Google also expands wildfire detection satellites through the Earth Fire Alliance partnership to help detect wildfires before they spread globally.
Google's 2026 Breakthroughs: AI for Time, Climate & Quality of Life
linkedin.comGoogle highlights 2026 AI breakthroughs including Gemini 3.1 Pro & Deep Think in Workspace, Groundsource for climate resilience using AI to help communities predict natural disasters, and WeatherNext Gen on Earth Engine with advanced diffusion-based ensemble weather models from DeepMind.
Ollama v0.34.3 Release - ChatGPT Desktop Support and Structured Output Improvements
github.comOllama v0.34.3 enables using Ollama models directly in ChatGPT Desktop on macOS, improves structured output performance on Apple Silicon, and adds support for OpenAI-compatible client tool search and response compaction features.
Trump signs voluntary AI accord with OpenAI, Anthropic, Google, Meta, xAI, Nvidia CEOs at White House
fortune.comTrump signed a voluntary AI accord at the White House with CEOs from major AI companies including OpenAI and Anthropic, addressing industry concerns about safety controls following recent security incidents.
Zuckerberg video launches Meta's new open-weight AI model
youtube.comMeta CEO Mark Zuckerberg announced the release of Muse Glimmer, a smaller open-weight foundation model designed to run agentic tasks on Mac or PC with a single graphics card. Meta is releasing more models as businesses grow wary of ballooning AI bills and cybersecurity incidents involving closed models from Anthropic, OpenAI and Meta.
QMFOL: Benchmarking Large Language Model Reasoning via Quantifiable Monadic First-Order Logic Test Case Generation - arXiv
arxiv.orgResearch paper introducing QMFOL, an automated framework that constructs formal logical structures and derives labeled reasoning tasks with controlled difficulty for scalable evaluation of model reasoning capabilities. The benchmark covers 960 configurations across logical and semantic factors with 2880 instances total. Published Jun 2026 (arXiv ID suggests June).
LLM Leaderboard & AI Model Benchmarks — September 2026 - BenchLM.ai
benchlm.aiComprehensive benchmark comparison table ranking 194+ models across 508 benchmarks including GPQA Diamond, ARC-Challenge, and BBH for reasoning tasks. GPT-6 Astra leads with a score of 88.69 on the composite metric, followed by Claude Opus 5.5 (86.98) and Claude Fable 5.1 (83). The leaderboard tracks models from OpenAI, Anthropic, Google, DeepSeek, Meta, Alibaba, and others with their reasoning capabilities. Published Sep 2026.
Best LLM Models 2026 Compared: Reasoning, Coding, Multimodal & Price - AI/ML API Blog
aimlapi.comAnalysis of spring 2026 LLM releases and their impact on reasoning benchmarks. DeepSeek V4 Pro emerges as a cost-effective option approaching frontier model quality, while closed models like GPT-5.5 and Claude Opus hold the edge on hardest academic reasoning tasks. The article covers benchmark performance across coding, scientific reasoning, knowledge, and long-context capabilities. Published Sep 2026 (based on search date).
2026 State and Federal AI Legislation Updates - Center for Democracy and Technology
cdt.orgThe 2026 legislative session showed continued focus on AI regulation, with most bills addressing chatbot use in various contexts. Connecticut passed comprehensive legislation including a regulatory sandbox and chatbot controls. The White House National Policy Framework for AI released March 2026 notes importance of combating AI-enabled scamming targeting vulnerable populations. Published Sep 30, 2026 (based on search date).
CAIDP Update: AI Policy News May 11 2026 | LinkedIn
linkedin.comCAIDP's weekly update tracking fast-moving developments in AI policy, privacy, and cybersecurity as of May 11, 2026. Published May 11, 2026.
Expert Predictions on What's at Stake in AI Policy in 2026 | TechPolicy.Press
techpolicy.pressExpert analysis of key AI policy issues in 2026, including the tug-of-war between states and federal government on preemption vs. state regulation, continued public skepticism about AI's ROI, concerns over civil liberties and misinformation, and calls for responsible development that protects people's rights. Published Sep 30, 2026 (based on search date).