Apple Lawsuit Alleges OpenAI Spy Ring, Model Wars Heat Up with GPT-5.6 and Muse Spark 1.1

Apple's bombshell lawsuit against OpenAI reveals a former engineer exploited a zero-day bug to steal confidential hardware specs after joining the rival firm. Simultaneously, the AI model wars intensify with OpenAI launching GPT-5.6 on AWS, Meta undercutting the market with a 75% cheaper Muse Spark 1.1 API, and Anthropic securing a $19 billion infrastructure deal while extending Claude Fable 5 access.

Written by AI from the sources it cites; it can be wrong — follow any citation to confirm.

11 sources1 min read

Apple's lawsuit details alleged OpenAI-led trade secret heist using rare security bug The 41-page complaint claims former Apple engineer Chang Liu exploited a zero-day authentication bug weeks after joining OpenAI, downloading sensitive hardware specs and circuit board designs. The suit alleges OpenAI conspired with current Apple employees, coaching them on how to avoid detection. 1 2 3

GPT-5.6 Sol, Terra, and Luna go live globally on Amazon Bedrock OpenAI's new frontier reasoning models are now generally available, matching their first-party pricing on AWS. Sol sets new benchmarks on coding (80% Coding Agent Index) and cybersecurity (73.5% ExploitBench). 4 5

Meta launches Muse Spark 1.1 developer API at roughly 75% below rivals' rates The new multimodal reasoning model marks Meta's pivot to a proprietary API business. Priced at roughly a quarter of OpenAI and Anthropic's flagship models, it features a 1 million token context window and strong agentic capabilities. 6 7

Anthropic signs $19 billion data center deal and extends Claude Fable 5 preview TeraWulf's 20-year Kentucky lease for Anthropic highlights the immense capital flowing into dedicated AI compute. In a direct response to GPT-5.6, Anthropic also extended developer access to its Claude Fable 5 model. 8 9

MIT Technology Review provides critical lens on Anthropic's 'J-space' AI discovery A detailed analysis examines what Anthropic's new window into Claude's internal reasoning does and does not reveal, cautioning against over-anthropomorphizing model behavior and misinterpreting mechanistic interpretability findings. 10

Activists stage protests at OpenAI, Anthropic, and Google DeepMind Coordinated demonstrations demanded a pause in frontier AI development, reflecting growing public concern over the rapid pace and safety of the industry's latest advances. 11