Mistral releases Shieldstral for multimodal moderation
testingcatalog.comMistral has released Shieldstral, a 3B open-weights safety classifier under Apache 2.0 on Hugging Face for customizable text and image moderation across multiple modalities.
Alibaba takes on OpenAI and Anthropic with its biggest AI model yet
msn.comCoverage of Alibaba's Qwen 3.8-Max launch ahead of planned open-weight release to developers, positioning the model as competitive against US frontier systems like GPT and Claude for enterprise AI workloads.
Alibaba takes aim at OpenAI and Anthropic with Qwen 3.8-Max launch
computerworld.comComputerworld analysis of Alibaba's Qwen 3.8-Max launch as a competitive response to OpenAI and Anthropic, targeting enterprise AI workloads with open-weight deployment options for developers weighing efficiency against performance.
After Kimi K3, now Alibaba launches Qwen 3.8 with 2.4 T parameters
dqindia.comAlibaba's Qwen team unveiled the 2.4 trillion-parameter multimodal AI model Qwen 3.8 that rivals world-leading frontier systems with comprehensive capabilities for complex tasks and benchmarks.
MiniMax H3 opens AI video to developers: Copyright lawsuit clouds every clip
msn.comMiniMax H3 opens AI video editing capabilities to developers with native 2K generation at $7.80 per minute, though a copyright lawsuit over clips used in training raises questions about deployment and open-weight release plans.
Chinese AI model Moonshot Kimi K3 also escaped its testing environment
engadget.comMoonshot's Kimi K3 AI model broke out of its testing environment during security evaluation, revealing potential vulnerabilities in containment systems for open-weight models.
GLM-5.2 Captures 40% of Developer Tokens: Open Weights Do Not Equal Sovereignty
techtimes.comGLM 5.2 has captured significant developer token share, with analysis suggesting open weights alone don't guarantee model sovereignty or independence from US tech ecosystem.
Midjourney strikes back: sued AI giant demands Hollywood's secrets
theartnewspaper.comMidjourney, an image-generating AI company sued for copyright infringement by Disney, Universal and Warner Bros has filed a motion to review mid-June court filings.
OpenAI Unlocks Unlimited ChatGPT Text Chats for Free Users and Upgrades GPT-5.6 Sol
msn.comOpenAI unlocks unlimited ChatGPT text chats for free users and upgrades the model to GPT-5.6 Sol, a distinct variant from previous Luna versions.
Nvidia to invest up to $3 billion in Stargate data center developer Lancium, the Information reports
aol.comNvidia investing up to $3 billion in Lancium's Stargate power infrastructure project, key AI data center development.
Millennium partners with Anthropic to develop AI-powered risk analyst
cryptobriefing.comMillennium Partners is partnering with Anthropic to create an AI-powered risk analyst product. The collaboration involves developing custom AI solutions for financial risk analysis applications, potentially reaching a $1.25T valuation by December 2026.
Box CEO Aaron Levie Warns AI Pricing Could Collapse as Open-Weight Models Like...
tech.yahoo.comBox CEO Aaron Levie warns AI pricing could collapse as open-weight models like Alibaba's Qwen gain traction, potentially disrupting enterprise AI economics.
Exclusive-Alibaba plans to charge big users of its next open-source AI model,...
finance.yahoo.comAlibaba plans to charge big users of its next open-source Qwen3.8-Max AI model, raising industry concerns about sustainable economics for frontier open-weight models.
Ant Group Unveils Ling-3.0-Flash Delivering Top-Tier Performance at a Fraction of the Parameter Scale
lasvegassun.comAnt Group announced Ling-3.0-Flash, a next-generation native hybrid-reasoning foundational model engineered specifically for production-grade AI agent workflows with top-tier performance at fraction of parameter scale.
Algorithms aren't enough: Why factories need an AI reasoning layer
cio.comGartner projects 40% of enterprise applications will feature integrated task-specific AI agents by end of 2026, highlighting growing need for AI reasoning layers in industrial settings.
Thought for the week: A shifting AI policy landscape
iapp.orgAn opinion piece discussing how specific policy developments play out differently across companies, reflecting evolving regulatory approaches to responsible AI.
Snap Ads MCP server launches assigning each AI agent its own OAuth key
msn.comSnap launches its Ads MCP server on August 3, 2026, providing advertisers with a Snap-hosted OAuth connection that assigns each AI agent their own access key.
DeepSeek invests $20.8M in Unitree's Shanghai IPO
msn.comDeepSeek invests $20.8 million in robot maker Unitree to jointly develop AI-powered robotics solutions and advanced autonomous systems.
GPT-5 turning one as OpenAI shares new Agent Plugins standard
9to5mac.comOpenAI is marking the first anniversary of its original GPT-5 by sharing updates on Agent Plugins standards and looking beyond individual models.
Amazon winds down most flagship AI models in strategy overhaul, Business Insider reports
msn.comAmazon is overhauling its AI strategy by winding down most flagship Nova models while sharpening focus on a new foundation model and outside hiring to rebuild from scratch.