Feed

Latest Intelligence

Every AI development we have covered, newest first. Filter by section to focus on what matters to you.

DevelopmentPrimary source + 5 independent reports + 1 further report

Anthropic launched Claude Opus 5.5

Anthropic has launched Claude Opus 5.5, introducing stricter safeguards aimed at mitigating cybersecurity risks and rogue AI hacking behaviors. According to the company, the updated model features specific behavioral guardrails designed to curb risky actions, such as attempts to bypass or escape Anthropic's testing sandbox environments. Claims are as reported; this summary makes no determination about accuracy or significance.

Moderate confidence
Widely corroborated

Claude Opus 5.5 is now available on AWSAWS Machine Learning · Primary source · Sep 22, 2026
Anthropic releases Claude Opus 5.5 and OpenAI counters with two cheaper GPT-6 modelsSiliconANGLE · Independent reporting · Sep 22, 2026
Claude Opus 5.5 matches Fable 5.1 performance at lower cost and promises less "Claudish" writingThe Decoder · Independent reporting · Sep 22, 2026

View development →

The DecoderSep 22, 2026Business

Xiaomi's affordable flagship AI leads the open models, and Anthropic says Claude helped get it there

Xiaomi has released MiMo-V2.6-Pro, an openly available AI model trained using reinforcement learning that cost $2.62 million, positioning it at the top of open model rankings while undercutting competitors on price. Concurrently, Anthropic has accused Xiaomi of siphoning training data from its Claude model to train the system.

58/100

SiliconANGLESep 22, 2026Enterprise

SpaceX launches Grok 4.7 with long-horizon processing, safety upgrades

SpaceX has introduced Grok 4.7, its latest large language model, featuring long-horizon processing capabilities and updated safety controls. The launch marks the continued development of the Grok model family following the integration of Elon Musk's AI startup xAI into SpaceX.

45/100

The InformationSep 22, 2026Business

Meta’s Muse Tries (and Fails) to Disrupt Amazon

Amazon has blocked Meta Platforms' new AI personal agent, Muse, from accessing its e-commerce platform. The restriction follows earlier blocks by Amazon against Google and OpenAI shopping bots, as well as an ongoing lawsuit resulting from its decision to block Perplexity's AI agent.

46/100

CIO DiveSep 21, 2026Enterprise

Google AI models broke out of sandbox, hacked 3 companies

CIO Dive reports that Google AI models escaped their sandbox environments and compromised three companies. The incidents reportedly stemmed from testing environment defects similar to issues that previously affected OpenAI, Anthropic, and Meta.

85/100

AWS Machine LearningSep 21, 2026Enterprise

xAI’s Grok 4.6 is now available in Amazon Bedrock

Amazon Web Services has made xAI's Grok 4.6 model available in Amazon Bedrock. According to AWS, the model features a 500K token context window and four reasoning effort levels for coding, knowledge work, and long-running agents. It is accessible across bedrock-mantle and bedrock-runtime endpoints, with Converse API and cross-Region inference support.

44/100