GPT-5.6 Pro undergoes stealth ChatGPT testing

// 2h agoNEWS

GPT-5.6 Pro undergoes stealth ChatGPT testing

Rumors and community leaks suggest OpenAI is stealth testing a next-generation AI model, GPT-5.6 Pro, directly within ChatGPT with a reasoning budget of 960 and native Playwright integration. While OpenAI has not officially announced the model, leaked details point to a December 2025 knowledge cutoff and native browser automation features ready for agentic web workflows.

// ANALYSIS

While stealth testing is a classic OpenAI play to generate hype, the combination of a 960 reasoning budget and native Playwright/browser-use integration signals a major shift from chat assistants to highly autonomous web agents.

–A reasoning budget of 960 suggests a dramatic increase in internal step-by-step thinking time, enabling the model to solve far more complex problems.
–Native integration of browser-use and Playwright indicates that OpenAI is aiming to deliver out-of-the-box browser-handling capabilities, paving the way for reliable agentic applications.
–The December 2025 knowledge cutoff confirms the model was trained very recently and is likely nearing a commercial rollout.

// TAGS

openaigpt-5.6-prochatgptleaksai-modelsweb-agentreasoning-budgetagent

DISCOVERED

2h ago

2026-06-20

PUBLISHED

2h ago

2026-06-20

RELEVANCE

9/ 10

AUTHOR

WorldofAI

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

OPEN SOURCE1h ago

Developer recreates Vercel keynote with open-slide

A developer has recreated Guillermo Rauch's Vercel Ship 26 London keynote presentation in three hours using open-slide and Cursor Composer. The project demonstrates the synergy between code-first slide frameworks and AI coding assistants, and the live reconstructed deck has been shared online.

NEWS1h ago

Greg Brockman endorses Codex App

OpenAI President Greg Brockman endorsed the Codex App, highlighting how the graphical agent orchestrator has replaced his traditional terminal-based development workflow. The desktop application enables developers to run multiple agents in parallel, utilize sandbox-based worktrees, and automate build-test-debug loops.

MODEL1h ago

Z.ai releases GLM-5.2 open-weights coding model

GLM-5.2 is a 744-billion-parameter open-weights Mixture-of-Experts model from Z.ai optimized for coding, frontend design, and long-context agentic reasoning. Trained using the open-source Slime RL framework, the model features a 1-million-token context window, speculative decoding, and adjustable reasoning effort modes to balance latency and quality in autonomous engineering loops.