engblogs

A technical overview of adding a secure personal-finance experience to ChatGPT, covering connected accounts via Plaid, a live dashboard for spending and balance insights, context-aware guidance powered by GPT-5.5, and privacy-first controls with configurable data settings and partner integrations like Intuit.

Databricks brings GPT-5.5 to enterprise agent workflows, delivering state-of-the-art 50% accuracy on OfficeQA Pro and a 46% reduction in errors over GPT-5.4 while enabling end-to-end parsing, retrieval, and orchestration through AI Unity Gateway, AgentBricks, and the Agent Supervisor API.

Leverage ChatGPT Work to consolidate scattered operational context—from trackers, dashboards, and notes—to produce decision-ready briefs, updates, and scenario models at speed, while teams retain final judgment.

Unveiling Gemini 3.5 Flash: a frontier-powered, agentic and coding-focused platform that accelerates long-horizon tasks with multimodal capabilities, safety safeguards, and scalable deployment across consumer, developer, and enterprise ecosystems.

A technical look at how evolving context and safety reasoning enable ChatGPT to recognize risk in sensitive conversations, use safety summaries across threads, and safely de-escalate or redirect users toward support.

Work with Codex from anywhere: how to use Codex across mobile and remote environments for live-state collaboration, secure cross-device control, and scalable automation.

Sea's Codex-driven approach reframes engineering as agentic, AI-native system orchestration at scale, intertwining automated testing, CI/CD, and architectural design to accelerate innovation across Southeast Asia.

New safety updates empower ChatGPT to detect evolving risk cues in sensitive conversations, using context-aware safety reasoning and safety summaries to de-escalate, refuse harmful details, and redirect users to support across single and multi-turn interactions.

Codex expands to the ChatGPT mobile app and remote environments, enabling secure, real-time collaboration, cross-device workflows, and automation as tasks run across laptops, devboxes, and remote machines.

Sea adopts Codex to enable agentic AI-driven software development at scale, transforming engineers into system orchestrators through automated, test-driven workflows across Southeast Asia.

A technical overview of safety updates enabling ChatGPT to recognize evolving risk signals in sensitive conversations through context, with safety summaries and de-escalation to safely refuse harmful requests and direct users to support.

Codex goes mobile to enable secure, cross-device collaboration and long-running work across laptops, remote environments, and enterprise workflows with live state, approvals, and automation.