Fortune reports Japan is emerging as a key market for Devin as aging software systems and a shrinking developer base sharpen demand for AI coding agents.

DeepSeek is reportedly doubling peak-hour API pricing for DeepSeek-V4 in July, a move that could push developers to rethink traffic timing and model mix.

A Yellow.com report says Meta told staff its internal ‘Watermelon’ AI model has caught GPT-5.5, signaling sharper competition in frontier models.

GLM-5.2 is being positioned as a cheaper rival to OpenAI and Anthropic, highlighting intensifying price pressure in enterprise AI models.

The UK AI Security Institute says standard benchmarks understate AI agent capability by capping compute, raising stakes for safety tests and buying decisions.

Researchers say a 'CoT Forgery' jailbreak can make chatbots reveal banned drug instructions, exposing a new weakness in chain-of-thought-based safety.

Coverage of Zhipu’s GLM-5.2 points to rising interest in lower-cost Chinese AI models, underscoring new pricing and deployment pressure in enterprise AI.

OpenAI, Anthropic and Meta are making fresh moves around Washington, model access and infrastructure, underscoring how AI competition now spans policy and cloud.

Argentina is weighing a framework for AI-run companies, but legal experts and policymakers still see humans as unavoidable for accountability.

The UN says AI capabilities are advancing faster than oversight, raising pressure on governments and companies to turn broad principles into enforceable rules.

Reports from BankInfoSecurity and TechRepublic highlight how AI agents are opening a new identity security gap for enterprises.

Bridgewater and Thinking Machines Lab say a tuned Qwen3-235B beat GPT and Claude on private finance tasks, highlighting the value of proprietary data.

Interfaze has released diffusion-gemma-asr-small, an open-source ASR model for six languages that tests diffusion decoding as a new path for speech AI.

OpenAI appears to be rolling out GPT-5.6 gradually, a cautious launch that highlights growing pressure around model reliability, cost and trust.

Jeff Bezos’ family office reportedly backed five AI startups in June, highlighting investor appetite across the AI stack despite limited deal details.

Insurance trade outlets are flagging a fresh wave of AI startup activity in insurance, highlighting investor interest but offering few verifiable details.

GitHub has made Kimi K2.7 Code generally available in GitHub Copilot, expanding model choice with an open-weight option for coding teams.

ElevenLabs is reportedly exploring a tender offer at a $22 billion valuation, signaling strong investor appetite for AI voice infrastructure.

Mark Zuckerberg told employees Meta’s AI agents are improving more slowly than he hoped, a notable reality check for a major platform bet.

Kuaishou’s Kling AI is reportedly raising nearly $3 billion at an $18 billion valuation, underscoring surging investor demand for video AI.