OpenAI and Broadcom unveil Jalapeño inference chip
OpenAI unveils its first in-house Intelligence Processor, built for LLM inference, with deployment scheduled to begin by the end of 2026.
SEARCH
Search English reports
English reports are listed here. AI-generated drafts are clearly labeled until an editor reviews them.
AI-generated · reviewedImpact 100/100
Google reportedly pushes Gemini 3.5 Pro to July
Google says Gemini 3.5 Pro is in internal use and due next month; Business Insider reports that its release has slipped to July 2026.
China issues network data security risk-assessment rules
China's cyberspace regulator publishes new risk-assessment rules for network data security, effective August 20, 2026.
OpenAI launches LifeSciBench for real-world life-science tasks
LifeSciBench uses 750 expert tasks to test evidence work, experiment design, translational judgment and scientific communication.
OpenAI promises two-part Codex reset after capacity incident
After resolving a model-capacity issue, OpenAI announces a global immediate reset and says a banked reset will follow.
xAI launches Grok Imagine Video 1.5
Grok Imagine Video 1.5 reaches general API availability with better motion, physics, audio and lip sync, plus a faster variant.
Z.ai releases GLM-5.2
Z.ai released GLM-5.2 for long-horizon agent tasks, with a stable one-million-token context window and MIT-licensed open weights.
Satya Nadella calls for human and token capital to compound together
Microsoft CEO Satya Nadella argues that companies should build learning loops between human and token capital instead of concentrating value in a few models.
Anthropic suspends access to Claude Fable 5
A U.S. export-control directive barred foreign nationals from Fable 5 and Mythos 5, leading Anthropic to suspend access for every customer because it could not verify nationality in real time.
OpenAI introduces banked usage resets for Codex
Eligible Plus and Pro users receive a Codex reset they can save for 30 days and redeem later, including one grant at launch.
Anthropic launches Claude Fable 5
Anthropic launched Claude Fable 5 as a Mythos-class model adapted for general use through a conservative runtime safety system.
NSPM-11 sets continuity requirements for military AI suppliers
NSPM-11 directs U.S. agencies to prevent commercial suppliers from disabling, degrading, blocking, or materially changing mission-dependent AI systems without government knowledge and approval.
OpenAI resets paid Codex limits after three reliability incidents
After three Codex reliability incidents within 24 hours, OpenAI resets usage limits across every paid plan.
Trump signs an executive order on advanced AI innovation and security
Executive Order 14409 creates a classified frontier-model benchmarking process and a voluntary early-access framework while explicitly rejecting mandatory licensing and preclearance.
OpenAI resets Codex limits at 5 million users
After Codex reaches 5 million users, OpenAI resets both weekly and hourly limits for every paid ChatGPT subscription.
Tencent launches WorkBuddy office agent worldwide
Tencent Cloud took its WorkBuddy office agent global, dispatching tasks and sub-agents in parallel from a single instruction to deliver ready-to-use work output.
Google launches Antigravity 2.0 agent development platform
Antigravity 2.0 uses a standalone desktop app to manage parallel agents, dynamic subagents, scheduled tasks and cross-product integrations.
Google DeepMind releases Gemini 3.5 Flash
Gemini 3.5 Flash is the first public model in the Gemini 3.5 series, aimed at agentic, coding, and tool-calling workflows.
OpenAI resets paid Codex limits after GPT-5.5 degradation fix
After fixing a GPT-5.5 capability degradation that lasted about 48 hours, OpenAI resets Codex limits for paid users.
Anthropic calls for tighter AI export controls and action on distillation
Anthropic published a policy piece arguing for closing chip and offshore-compute loopholes, restricting frontier-model access, and curbing model-distillation attacks.