How to use GPT-5.6

2026-05-14 · Source: Ben's Bites · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Robotics & Autonomous Systems · Depth: Intermediate, medium

Summary

OpenAI has released its GPT-5.6 models, integrating the ChatGPT macOS app and Codex app into a new "ChatGPT Work" mode, alongside a "ChatGPT Sites" plugin for hosted websites. The GPT-5.6 series includes Luna, Terra, and Sol models, each offering five thinking levels (light to max) and an "Ultra mode" for subagent use. Higher thinking levels, especially Ultra mode, consume usage limits rapidly. OpenAI temporarily removed the 5-hour usage limit during bug fixes. Sol excels at UI tasks and writing, Terra offers minor improvements over 5.5 and is steerable, while Luna requires clearly defined prompts. A notable feature is the Codex app's "Computer Use" capability, allowing models like Sol medium/high to self-drive cursors and interact with applications.

Key takeaway

For AI Engineers evaluating new model deployments, the GPT-5.6 series offers distinct capabilities and usage considerations. You should strategically select models like Sol for UI or Luna for daily tasks, carefully managing thinking levels to optimize performance and control API costs, especially given the temporary removal of usage limits. Experiment with the "Computer Use" feature in the Codex app to explore new automation possibilities.

Key insights

OpenAI's GPT-5.6 introduces tiered models and thinking levels, demanding strategic usage to manage resource consumption and optimize task performance.

Principles

• Higher AI thinking levels increase resource consumption.
• Model performance varies by task and prompt clarity.
• Reference data improves UI generation quality.

Method

Match GPT-5.6 models and thinking levels to task complexity: Sol medium for creativity, Luna xhigh for daily productivity, and background agents for harder tasks.

In practice

• Use Sol medium/high for "Computer Use" to automate cursor interaction.
• Provide UI references to Sol for improved interface generation.
• Define Luna prompts clearly to avoid ambiguity failures.

Topics

• GPT-5.6 Models
• AI Agent Systems
• Model Usage Optimization
• Computer Control
• LLM Deployment
• API Cost Management

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Engineer, Machine Learning Engineer, AI Product Manager

Related on AIssential

• OpenAI Launches GPT-5.6 Family with Sol, Terra, and Luna for Flexible AI Choices — OpenAI's GPT-5.6 family offers tiered AI models (Sol, Terra, Luna) balancing intelligence, speed, and cost for diverse applications.
• How to Work Effectively with GPT-5.6 — GPT-5.6 offers incremental improvements over GPT-5.5, but effective use requires careful management of model size, reasoning levels, and tool access.
• The new GPT-5.6 family: Luna, Terra, Sol — OpenAI's GPT-5.6 models prioritize agentic efficiency and introduce advanced API features for complex workflows.
• GPT 5.6 is HERE! but you can't use it! — Government restrictions are delaying broad access to OpenAI's new, highly performant, and cost-effective GPT 5.6 model family.
• [AINews] OpenAI launches GPT 5.6 Sol/Terra/Luna, Codex becomes ChatGPT superapp — OpenAI's GPT-5.6 models offer superior performance and cost-efficiency, integrated into an expanded agentic work platform.
• Quoting OpenAI — OpenAI is launching the GPT-5.6 series, offering flagship, balanced, and cost-effective models with improved performance and pricing.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Ben's Bites.