WaperOpenAI introduced GPT-6 Cyber, expanded its security offering, and announced new enterprise, API, and leadership developments.
OpenAI tracker6 items
New5
OpenAI and JPMorganChase announce strategic AI partnership
OpenAI and JPMorganChase will explore using advanced AI across engineering, cybersecurity, and employee and customer applications. Greg Brockman highlighted cooperation on protecting the financial system.
Why it matters: The partnership signals deeper adoption of OpenAI systems in a major financial institution, especially for cybersecurity and enterprise workflows.
OpenAI makes GPT-Live outbound telephony API generally available
The GPT-Live outbound telephony API is now generally available, allowing voice agents to place calls through any SIP trunk provider, including Telnyx or Twilio. Calls support HD Voice and keypad events.
Why it matters: Developers can now deploy production voice agents that make phone calls and handle DTMF interactions through existing telephony infrastructure.
Claude Code product lead Catherine Wu joins OpenAI
Catherine Wu, Anthropic’s founding product lead for Claude Code, is joining OpenAI to lead product management for Core Products and work with Codex lead Thibault Sottiaux.
Why it matters: The hire strengthens OpenAI’s product leadership as competition for coding-agent talent and developer workflows intensifies.
OpenAI Python SDK 3.46.0 adds regional residency and image controls
The release adds API residency options for Australia, India, and Japan; GPT-6 Rosalind to the model catalog; GIF and SVG output controls; and expanded agent, image, and audio options.
Why it matters: The update gives developers more deployment-locality choices and finer control over newer image and agent capabilities.
OpenAI introduces Private Safety Processing for API requests
Private Safety Processing lets OpenAI review related API requests without retaining prompts or responses. Records remain in the customer’s environment, where an automated safety agent reviews content without OpenAI personnel access.
Why it matters: The option may make OpenAI’s safety review more acceptable for developers handling sensitive or regulated data.
Follow-ups1
OpenAI launches GPT-6 Cyber for security research and defense
Mozilla engineer Brian Grinstead says GPT-6 Cyber found novel Firefox vulnerabilities, generated reproducible test cases and patches, and helped engineers quickly ship fixes to millions of users. He described the tool as complementing experienced researchers by uncovering subtle bugs and strengthening Firefox’s layered defenses.
Why it matters: This provides concrete evidence of GPT-6 Cyber’s use in real-world vulnerability discovery and remediation at scale.