ChatGPT logo

ChatGPT: Choose the Plan, Mode and Review Step

AI workspace for chat, research, files, coding and multi-step work, with plan-specific models, speeds and limits.

AI Models: LLMs, Multimodal Systems, and More
Visit ChatGPT → Join Discussion
WHATAI LATEST · OCT 5, 2026

Pro 500 Makes Speed a Separate Buying Decision

Astra Ultrafast can reduce response latency in Work and Codex, but its US$500 monthly price and higher allowance consumption make measurement essential.

By WhatAI Editorial ·

Pro now has three different prices

OpenAI now lists Pro 100 at US$100 per month, Pro 200 at US$200 per month and Pro 500 at US$500 per month. The tiers differ in included usage, and Pro 500 has the highest allowance. This makes the upgrade question more specific than whether a user wants a general premium plan.

Ultrafast is a speed tier, not a quality guarantee

Pro 500 is the only individual Pro tier that includes Astra Ultrafast in ChatGPT Work and Codex. OpenAI describes it as a lower-latency service tier. The same guidance says Ultrafast uses included allowance at eight times the Standard rate, and that the rate is not a promise of a particular speed increase.

ChatGPT home screen showing the chat input, sidebar navigation and account options

Measure wait time against allowance use

A useful test is one repeated task with fixed inputs and a clear acceptance check. Run it using the modes available to the account, then record elapsed time, allowance consumption, corrections and the final output. A faster first response is not a saving if review or rework grows.

WhatAI view

Pro 500 has a credible role for people whose repeated Astra workloads are genuinely delayed by response time. It is harder to justify for occasional use, vague experimentation or tasks where Standard or Fast already meets the deadline. Buy for a measured bottleneck, and check the model picker, allowance and credit rules on the real account before upgrading.

ℹ️

WhatAI Decision Box

✓
Best for:

People who want one workspace for chat, research, files, coding and connected tasks, and teams that can measure whether higher allowances or faster Astra responses improve a repeated workflow.

✗
Not for:

Work that requires guaranteed correctness, fully offline operation, invisible external actions or a US$500 speed tier without a measured time-saving case.

⇆ Often compared with

ℹ️ WhatAI Field Note

  • Benchmark one repeated Work or Codex task with fixed inputs and acceptance checks. Record elapsed time, allowance use, corrections and final quality.
  • Check the actual plan, model, speed mode, region and workspace permissions on the account where the task will run.

ChatGPT spans chat, research, files, connected apps, Work and Codex. The practical choice is now partly about capacity and speed: Free, Go and Plus cover broad everyday use, while Pro 100, Pro 200 and Pro 500 separate higher allowances from Astra Ultrafast access.

Pro 500 adds Ultrafast, not automatic value

Pro 500 costs US$500 per month and is the only individual Pro tier with Astra Ultrafast in Work and Codex. OpenAI says Ultrafast uses included allowance at eight times the Standard rate, so faster responses should be tested against the real task and its total correction time.

Keep permissions, evidence and outputs visible

A higher tier does not remove model limits, connected-app permissions or the need to inspect sources, calculations, code and exported files. Test one repeated workflow, record elapsed time and allowance use, and compare the corrected result with a lower tier.

About ChatGPT

ChatGPT is OpenAI's general-purpose AI workspace for conversation, research, files, images, coding and reviewable multi-step work. Chat suits questions and focused drafts, Work carries defined deliverables, and Codex handles software tasks. Paid access now spans several individual and team tiers, including Pro 100, Pro 200 and Pro 500. Pro 500 adds Astra Ultrafast in Work and Codex, but speed, model access, limits and permissions still depend on the plan, surface and workspace.

Use Cases

Develop an idea into a brief and an editable presentationResearch a decision with traceable source evidenceAnalyse source files and prepare a reportDraft and revise content against examples and constraintsCarry out a bounded software task in CodexCompare Standard, Fast and Ultrafast on a repeated Work or Codex taskTalk through a task and continue it in text after Voice endsUse approved plugins during a conversationScan paper notes into a PDF on supported iOS devicesInvestigate connected business data and build a dashboard

Key Features

  • ✓ Chat for questions, brainstorming and focused drafts
  • ✓ ChatGPT Work for defined multi-step deliverables
  • ✓ Codex for development and technical workflows
  • ✓ GPT-6 Astra in eligible ChatGPT, Work and Codex plans
  • ✓ Astra Ultrafast in Work and Codex on Pro 500 and eligible workspaces
  • ✓ Voice on supported web, mobile and desktop surfaces
  • ✓ Web search with source links
  • ✓ Documents, spreadsheets, presentations and PDFs
  • ✓ Image creation and revision where available
  • ✓ File uploads and data analysis with plan-specific limits
  • ✓ Plugins and connected apps subject to identity and permissions
  • ✓ Finances for eligible accounts and regions
  • ✓ Deep research with plan-specific limits
  • ✓ Memory and projects where available

Pricing

Free

$0

  • • Entry-level access; individual tool limits apply

Go

US$8/month

  • • Lower-cost paid tier; verify local checkout, ads eligibility and current feature limits

Plus

US$20/month

  • • Expanded models, tools and Work access; usage limits apply

Pro 100

US$100/month

  • • Pro models and higher Work and Codex allowance
  • • Astra Ultrafast is not included

Pro 200

US$200/month

  • • More included usage than Pro 100
  • • Astra Ultrafast is not included

Pro 500

US$500/month

  • • Highest included usage among Pro tiers
  • • Astra Ultrafast in Work and Codex
  • • Ultrafast uses allowance at a higher rate

Business Standard

US$25/user monthly or US$20/user/month billed annually

  • • At least two paid seats
  • • Managed workspace with ChatGPT, Work and Codex

Business Premium

US$125/user monthly or US$100/user/month billed annually

  • • At least two paid seats across Standard and Premium
  • • Higher included usage than Standard

Enterprise

Custom

  • • Contracted administration, access, support and billing terms
  • • Model rollout and speed modes depend on the agreement and workspace settings

Pricing varies by plan and region — see current pricing.

Plan features change — last updated: 2026-10-05.

Details

Categories: AI Models: LLMs, Multimodal Systems, and MoreAI Writing & ContentAI, Coding and DevelopmentAgents & AutomationCommunicationDesign & CreativeEducationFrontier LLMs & Chat ModelsGeneralResearch & Knowledge Work
Skill Level: beginner
Access Methods: browser, desktop, ios, android

Tags

chatgptopenaiai assistantchatgpt workcodexgpt-6-astraastra ultrafastpro 500voice aiconnected appsresearchfile creation

ChatGPT Community Discussions

Explore community discussions. Ask and answer questions on ChatGPT to grow and learn together.

WhatAI Editorial · ChatGPT AI Models: LLMs, Multimodal Systems, and More

ChatGPT Images 2.5: can you change one detail without rebuilding the picture?

The frustrating part of an AI-generated image often comes after the first good result: you ask for a small change and something you wanted to keep changes too. For a shop owner making product artwork or a creator preparing a series of posts, that can turn a quick edit into another round of checking. OpenAI’s September 8 announcement for ChatGPT Images 2.5 puts more emphasis on this problem. The company says it improves consistency across repeated edits and adds comments placed directly on an image for more focused changes. It also introduces Sketch as a visual reference tool. OpenAI says the release spans all ChatGPT tiers; that does not establish identical usage allowances for every account. Here is OpenAI’s short introduction, useful as a starting point for what the company is promising: A practical test would be to take one image you have permission to use and ask for three… Read full discussion →
♥ 3 💬 0 👁 6 Reply →
phoenix_campbell · ChatGPT AI Models: LLMs, Multimodal Systems, and More

Deep Research and Thinking mode are the two ChatGPT features actually worth learning properly

I had been using the same ChatGPT settings for everything for months. After working through I went back and rebuilt how I use it for two specific types of work. Deep Research producing comprehensive cited reports rather than search results changed what questions I am willing to try to answer in a session. A research task that previously required hours of manual synthesis across multiple sources now produces a structured report in a timeframe I can actually use. The quality ceiling depends on question complexity. The floor has risen significantly. Thinking mode for complex analytical work is the setting worth using on multi-variable problems where the reasoning chain is what produces the right answer. Strategy decisions, technical architecture, pricing scenarios with competing considerations. The extra time is worth it when shallow reasoning produces the wrong answer. The Canva and Adobe Photoshop integrations available through the same interface as text generation… Read full discussion →
♥ 1 💬 2 👁 8 View 2 replies →
ethanG · ChatGPT AI Models: LLMs, Multimodal Systems, and More

The ChatGPT app integrations with Canva and Photoshop are worth testing if you work across both

I covered Deep Research and Thinking mode separately. The same practical demonstration video, also covers the app integrations which deserve their own attention for anyone regularly working in Canva or Photoshop. The Canva integration working from within a ChatGPT conversation means you can describe a visual need, generate the asset through Canva and continue working in the same conversation without switching tabs. For workflows that move between research, writing and visual output frequently the context switch reduction is real and cumulative across a working day. The Adobe Photoshop integration is the professional creative tool connection that changes which users can work fluidly with ChatGPT as a creative collaborator rather than a text-only tool. Image generation with aspect ratio controls in the same interface as text generation and research means a single ChatGPT session can cover the research brief, the written content and the visual output. Not all at production quality… Read full discussion →
♥ 1 💬 2 👁 7 View 2 replies →
gabriel107 · ChatGPT AI Models: LLMs, Multimodal Systems, and More

ChatGPT's customisable personas changed how I start conversations and they are being widely overlooked

The newest ChatGPT update walkthrough and there are a few things in it worth paying attention to even if you use ChatGPT regularly enough to think you already know what changed. The customisable personas via the Voice tab are the feature most coverage glosses over. Being able to adjust the AI's tone and personality means a persona calibrated for technical documentation produces meaningfully different output from one set for casual copy, even on identical prompts. For people doing varied writing work that is not a minor tweak. The sidebar redesign is genuinely useful rather than cosmetic. Managing multiple concurrent chat threads through the previous interface was friction that built up slowly. A dedicated sidebar that makes navigation between threads fast changes how you work when several things are running simultaneously. The improved context retention across a long conversation is real and noticeable in practice. Earlier versions had a drift pattern… Read full discussion →
♥ 1 💬 2 👁 12 View 2 replies →
reeseturner · ChatGPT AI Models: LLMs, Multimodal Systems, and More

GPT-5 through GPT-5.4 changed the model landscape and this breakdown explains what actually matters

Version numbers are meaningless without understanding what the actual capability differences are. The model evolution video traces the changes from GPT-4o through GPT-5 to GPT-5.4 and the GPT-5 to GPT-5.4 jump is the more significant one for daily use. GPT-5 introducing Instant and Thinking modes was the first meaningful mode differentiation. The discipline of choosing the right mode for the task rather than defaulting to one setting for everything is what separates efficient use from treating a more capable model like a faster version of the previous one. GPT-5.4 with million-token context windows is where the scope of a single session changes in a way you can actually feel. The Atlas Browser combining integrated web browsing with persistent memory is the research workflow change that makes GPT-5.4 feel categorically different from GPT-5 for knowledge work. Canvas Mode for collaborative text and code editing is the working document environment that changes… Read full discussion →
♥ 1 💬 2 👁 12 View 2 replies →
View All ChatGPT Discussions
Gallery

ChatGPT Showcase

5 items
ChatGPT Images 2.5: can you change one detail without rebuilding the picture?

ChatGPT Images 2.5: can you change one detail without rebuilding the picture?

WhatAI Editorial
Deep Research and Thinking mode are the two ChatGPT features actually worth learning properly

Deep Research and Thinking mode are the two ChatGPT features actually worth learning properly

phoenix_campbell
The ChatGPT app integrations with Canva and Photoshop are worth testing if you work across both

The ChatGPT app integrations with Canva and Photoshop are worth testing if you work across both

ethanG
ChatGPT's customisable personas changed how I start conversations and they are being widely overlooked

ChatGPT's customisable personas changed how I start conversations and they are being widely overlooked

gabriel107

Recommended Watch

See ChatGPT image creation in action

Recommended Watch

OpenAI’s Introducing ChatGPT Images 2.0 demonstration gives you a visual starting point for exploring image creation in ChatGPT. It covers an earlier release, so use it for workflow context alongside the Images 2.5 update above, rather than as evidence of current model availability.

👍 👎

ChatGPT Pros & Cons

Workflow breadth

👍 Pro

Moves between chat, files, research, Work and Codex.

👎 Con

Specialist tools may still offer finer final control.

Plan choice

👍 Pro

Several individual and team tiers cover different capacity needs.

👎 Con

The expanded pricing ladder makes direct comparison more complex.

Ultrafast

👍 Pro

Can reduce latency for eligible Astra work in Work and Codex.

👎 Con

Pro 500 costs US$500 per month and Ultrafast consumes allowance faster.

Connected work

👍 Pro

Can use approved apps and plugins inside a workflow.

👎 Con

Identity, permissions and approval-required actions must be checked.

Research

👍 Pro

Web search can provide inspectable source links.

👎 Con

A citation does not guarantee that the exact claim is supported.

How to Get Results with ChatGPT: Step-by-Step Workflow

  1. Define the repeated task

    Choose a real Work or Codex job with fixed inputs, output and acceptance criteria.

  2. Record the account and mode

    Note the plan, model, speed mode, workspace and any connected data permissions.

  3. Run the same task

    Use the same brief and permitted sources so the comparison tests speed rather than a different job.

  4. Measure more than latency

    Record elapsed time, allowance use, corrections, tool failures and final quality.

  5. Review evidence and actions

    Check sources, calculations, code changes and any approval-required action before accepting the result.

  6. Decide from the bottleneck

    Upgrade only if the time saved on repeated work outweighs the price and higher allowance consumption.

ChatGPT Gotchas and Limits to Know Before You Start

  • Pro 100, Pro 200 and Pro 500 have different included usage.
  • Astra Ultrafast is included with Pro 500, not Pro 100 or Pro 200.
  • Buying credits on Pro 100 or Pro 200 does not unlock Ultrafast at launch.
  • Ultrafast uses included allowance at eight times the Standard rate.
  • Allowance consumption rates do not guarantee a particular speed increase.
  • Model and speed options depend on the plan, workspace settings and current rollout.
  • Connected apps do not grant access beyond the signed-in user's existing permissions.
  • API-key usage is billed separately from ChatGPT subscription allowances.
  • Sources, exported files and generated code still require review.
  • This page offers evaluation guidance, not a WhatAI hands-on benchmark.

Which ChatGPT Feature Fits Your Use Case

Feature Good for Common mistake Fix
Astra Ultrafast Reducing latency on eligible repeated Work or Codex tasks Assuming higher allowance consumption guarantees a proportional time saving Benchmark elapsed time, corrections and allowance use on the real task.
ChatGPT Work Carrying a defined task to a reviewable file or result Starting without a clear acceptance test Specify the audience, output and checks before the task begins.
Web search Current evidence and source discovery Trusting a cited claim without opening the source Check the source date and whether it supports the exact statement.
Connected apps Using permitted account context in a workflow Using the wrong connected identity Confirm the account, permissions and proposed action.
Codex Bounded software changes and technical checks Treating generated code as verified Review the diff and run proportionate tests.

Starter Prompts for ChatGPT

Run this representative Work task with the attached sources. Produce the requested document, cite each important claim, and report elapsed time, corrections and any unresolved issue.
Compare the speed modes available on this account using the same coding task. Record elapsed time, allowance change, test results and any difference in corrections. Do not change the repository until I approve the proposed diff.
Research the claims in this draft. Return a table of claim, supporting source, source date and unresolved uncertainty. Do not replace missing evidence with confident wording.
Using the approved project files, prepare a one-page meeting brief with decisions, risks and open questions. Cite the source for each important claim.
Review this code change against the issue and tests. Explain the proposed diff before editing, then run the relevant checks and report any unverified behaviour.

WhatAI view: pay for a measured bottleneck

Pro 500 is easiest to justify when waiting on repeated Astra work is a costly bottleneck and the saved time survives review. Run the same representative Work or Codex task in the modes available to your account, then compare elapsed time, allowance use, corrections and final quality. If Standard or Fast already meets the deadline, the higher monthly price may buy little practical value. If Ultrafast shortens a frequent, expensive workflow without increasing corrections, the upgrade has a clearer case. Treat speed mode as an operating choice, not a quality guarantee.

ChatGPT — Frequently Asked Questions

What are the current ChatGPT Pro tiers?

OpenAI lists Pro 100 at US$100 per month, Pro 200 at US$200 per month and Pro 500 at US$500 per month. Included usage differs by tier.

Which individual plan includes Astra Ultrafast?

At launch, Astra Ultrafast in ChatGPT Work and Codex is included only with Pro 500. Buying credits on Pro 100 or Pro 200 does not unlock it.

Does Ultrafast guarantee a better answer?

No. It is a lower-latency service tier for GPT-6 Astra. OpenAI describes allowance consumption rates, not a guaranteed speed increase or higher output quality.

How quickly does Ultrafast use allowance?

OpenAI says GPT-6 Astra Ultrafast uses included allowance at eight times the Standard rate. Actual usage also depends on context, reasoning, tool use, retrieval and the task.

How much does ChatGPT Plus cost?

OpenAI lists ChatGPT Plus at US$20 per month. API usage is separate and billed independently.

How does ChatGPT Business pricing work?

OpenAI lists Standard seats at US$25 per user monthly or US$20 per user per month billed annually, and Premium seats at US$125 monthly or US$100 per month billed annually. A Business workspace requires at least two paid seats.

Should I use Chat, Work or Codex?

Use Chat for questions and focused drafts, Work for research and reviewable documents or data tasks, and Codex for software changes and repository work. Availability depends on the plan and workspace.

How should I test a plan upgrade?

Choose a repeated task with fixed acceptance criteria. Compare elapsed time, allowance use, corrections, source quality and the final output before deciding whether the higher tier solves a real bottleneck.

Related AI Models: LLMs, Multimodal Systems, and More Tools

5 tools
Claude logo

Claude

$0–$200/mo

 Google Gemini logo

Google Gemini

$0–$199.99/mo

 Perplexity logo

Perplexity

$0/mo – Custom

 Grok AI logo

Grok AI

$0 – Custom

 OpenAI Codex logo

OpenAI Codex

$0/mo – Custom

Explore the Network

People discussing ChatGPT also discuss...

Alternatives to ChatGPT

Claude Claude $0–$200/mo Compare Google Gemini Google Gemini $0–$199.99/mo Compare Perplexity Perplexity $0/mo – Custom Compare Grok AI Grok AI $0 – Custom Compare

Pairs well with ChatGPT

Sources & References

  1. ChatGPT release notes ↗
  2. ChatGPT Voice ↗
  3. Connected apps in ChatGPT ↗
  4. ChatGPT pricing ↗
  5. ChatGPT Business and Enterprise credit rate card ↗
  6. ChatGPT Go availability and pricing ↗
  7. Data agent in ChatGPT Work ↗
  8. GPT-6 Astra and the expansion of AI work ↗
  9. What is ChatGPT ↗
  10. ChatGPT for Financial Services ↗
  11. About ChatGPT Pro tiers ↗
  12. ChatGPT Work and Codex ↗
  13. ChatGPT Business general FAQ ↗
  14. What is ChatGPT Plus? ↗

Try ChatGPT

Visit the official website to get started with ChatGPT today.

Visit ChatGPT →

Explore More

More AI Models: LLMs, Multimodal Systems, and More Tools

Browse similar AI tools in this category

Compare AI Tools

Side-by-side comparison of features

Community Forum

Discuss ChatGPT with other users