ChatGPT

ChatGPT is a conversational AI assistant that uses large language models (LLMs) trained on public, partner‑provided, and human‑generated data, combining text, images, audio, and video sources to understand and generate human‑like responses.

It operates using transformer architecture where tokens (words, parts, punctuation) are processed to predict the next likely token. This underpins tasks such as writing, coding, summarizing, translating, reasoning, and multi‑step workflows.

Recent model upgrades include GPT‑5.6 (Sol, Terra, Luna variants), introduced in July 2026, enhancing reasoning and depth, including a new “ChatGPT Work” mode tailored for complex, multi‑step tasks.

• Custom GPTs: Allows creation of tailored assistants with instructions, knowledge, tools, and APIs—enhancing personalization and task-specific utility. Available to all users, creation/editing requires a paid subscription.
• Memory & Memory Sources: Enables more personalized and continuous interaction by drawing on past chats, saved memories, uploaded files, and connected Gmail data—visible and editable as sources under responses.
• ChatGPT Work Mode: Introduced with GPT‑5.6 release, it supports multi‑step task workflows across desktop, web, and mobile, improving productivity for complex tasks.
• GPT‑5.6 Model Variants: Latest generation includes Sol, Terra, and Luna models offering varying reasoning depth and compute tradeoffs, accessible across subscription tiers.
• Cross‑Platform Desktop Experience: Unified interface enables switching between Chat and Work, syncing projects and conversations across devices, with improved app navigation and stability.

• Content Creators & Writers: Use Work mode and GPT‑5.6 reasoning for drafting, editing, and iterative writing tasks with memory‑driven personalization.
• Developers & Coders: Custom GPTs and Codex integration streamline coding, project scaffolding, and multi‑step automation within ChatGPT Work.
• Knowledge Workers & Analysts: Leverage deep research tools, memory, and multi‑step workflows to analyze data, synthesize information, and execute compound tasks.
• Teams & Enterprises: Use Team or Enterprise plans to enforce security, admin control, SSO, and access advanced GPTs for organizational workflows.

• Account setup: User signs in on web, desktop, or mobile—selects Free, Go, Plus, Pro, Team, or Enterprise tier.
• Choose mode: In desktop or app, switch between Chat or Work mode via the global toggle.
• Input and context: User enters prompt, optionally uploads files, connects Gmail or tools (on eligible plans), or opens a Project.
• Model selection: System routes request to a model variant (e.g., GPT‑5.2 Instant, GPT‑5.5, or GPT‑5.6 Sol/Terra/Luna) depending on tier and task complexity.
• Response generation: Model generates output token by token, utilizing memory, custom GPT instructions, and connected files to tailor results.
• Memory & feedback: Responses include visible source references; user can edit or remove memory to refine personalization.
• Continue workflow: Conversations and Projects sync across devices; users can manage tasks, projects, and agents across sessions.

ChatGPT uses tiered subscription pricing with multiple consumer and business levels.

Consumer tiers include:

• Free (limited access to GPT‑4.1 mini, GPT‑4o, limited memory/features)
• Go (mid‑tier: expanded usage, GPT‑5.2 Instant, longer context/window)
• Plus (deep reasoning access, GPT‑4.1, GPT‑4.5 research preview, custom GPTs)
• Pro (highest access: unlimited reasoning models, GPT‑4o, advanced voice/video, Codex agent previews)

Business and Enterprise tiers offer seat‑based pricing with security, admin controls, GPT‑5.6 Sol/Terra/Luna access, ChatGPT Work, plugins, SSO, and data privacy. Team is billed per user/month; Enterprise custom‑quoted for scale. Free trials/freemium available in Free and Go tiers.