Navigation
Explore
Explore
OpenAI's general-purpose AI assistant powering text, image, voice, code, and agentic workflows across six subscription tiers from free to enterprise.
GPT-5.4 (limited) and GPT-4o mini fallback at limits.
Limited daily messages, uploads, and image generations; GPT-5.4 access throttled; ads shown in the US.
Content Creation
8.5GPT-5.4 supports long-form writing, image generation, voice narration, and canvas-based editing in a single interface; Plus plan includes Sora for video generation.
Coding
8.5Codex agent (Plus and above) autonomously completes multi-step coding tasks with GitHub integration and automated PR creation; GPT-5.4 scores ~57.7% on SWE-bench Pro.
Research
8.0Deep Research mode (10 runs/month on Plus) generates multi-source structured reports with citations; Pro users receive higher Deep Research limits.
Data Analysis
8.0Native code execution and file uploads enable spreadsheet analysis, chart generation, and data summarization; Plus tier unlocks full data analysis with larger file uploads.
Personal Productivity
8.5Tasks, Projects, Memory, and 60+ app integrations (Slack, Google Drive, GitHub, SharePoint) centralise multi-tool workflows; Business plan includes shared projects and custom workspace GPTs.
ChatGPT is OpenAI's consumer and enterprise AI platform, built on the GPT-5.4 model family launched March 5, 2026. Six tiers serve different user profiles: Free (GPT-5.4 limited, ad-supported in the US from February 2026), Go ($8/month, expanded limits but ad-supported), Plus ($20/month, GPT-5.4 Thinking, Deep Research, Sora, Codex agent, Agent Mode), Pro ($200/month, unlimited GPT-5.4 Pro with guaranteed compute), Business ($25/user/month, workspace privacy, SAML SSO), and Enterprise (custom). GPT-5.4 unifies reasoning, coding, native computer use, and image generation. Primary limitation: advanced features like GPT-5.4 Thinking, Sora, and Codex agent require Plus ($20/month) or above; Free and Go tiers include advertising.
Pricing
| Plan | Model | Usage Limits | Price |
|---|---|---|---|
| FreeFREE | GPT-5.4 (limited) and GPT-4o mini fallback at limits. | Limited daily messages, uploads, and image generations; GPT-5.4 access throttled; ads shown in the US. | Free |
| Go | GPT-5.4 standard; no Thinking or Pro variants. | More capacity than Free; no Ads-Free option; excludes GPT-5.4 Thinking, Sora, Codex, Agent Mode, Deep Research. | $8/mo |
| Plus | GPT-5.4 Thinking (80 msgs/3h window) and GPT-5.4 standard. | GPT-5.4 Thinking at 80 messages per 3-hour window; 10 Deep Research runs/month; ad-free; 256K reasoning context window. | $20/mo |
| Pro | GPT-5.4 Pro (unlimited) and GPT-5.4 Thinking (unlimited). | Unlimited GPT-5.4 Pro; 400K reasoning context window; dedicated GPU slice; priority capacity. | $200/mo |
| Business | GPT-5.4 standard with higher limits; GPT-5.3 Instant available (admin-enabled). | GPT-5.4 with increased limits; workspace privacy; no training on data by default; SAML SSO and MFA; 256K reasoning context window. | $25/user/mo, annual |
| Enterprise | GPT-5.4, GPT-5.3 Instant, GPT-5.4 Thinking/Pro (admin role-based access control); retired models (GPT-4o) available until April 3, 2026. | Unlimited GPT-5.4; advanced security; SCIM provisioning; audit logs; 250-page context window equivalent. | custom |
GPT-5.4 (limited) and GPT-4o mini fallback at limits.
Limited daily messages, uploads, and image generations; GPT-5.4 access throttled; ads shown in the US.
Plus at $20/month bundles text, image (GPT-5.4 image), and video (Sora) generation in one subscription, replacing multiple separate tools; Canvas enables iterative document editing without re-prompting.
Codex agent on Plus autonomously implements multi-file code changes, runs tests, and creates pull requests from GitHub issues; GPT-5.4's 57.7% SWE-bench Pro score is competitive with frontier coding models.
Deep Research mode synthesises multi-source reports with structured citations; Plus users get 10 Deep Research runs/month; Pro users get higher limits suitable for daily intensive research workflows.
Business plan at $25/user/month includes workspace privacy, no training on data by default, SAML SSO, and 60+ native app integrations; Enterprise adds SCIM, audit logs, and advanced compliance for regulated organisations.
Consider These Instead
Choose Claude Pro ($20/month) if your primary workflows are long-document analysis, extended reasoning sessions, or code review — Claude Opus 4.6's 1M context window and 75.6% SWE-bench score are competitive with GPT-5.4 for these specific tasks. Choose Perplexity AI Pro ($20/month) if real-time, citation-native web research is the primary use case — Perplexity's source-grounded architecture provides stronger fact-traceability than ChatGPT's Deep Research for high-frequency research workflows.
