ChatGPT Plan Changes with GPT-5.6: Free, Plus, Pro, Former Team, Business, and Enterprise Access Compared

GPT-5.6 changes the boundaries between ChatGPT subscriptions by reserving its flagship Sol reasoning model for eligible paid plans, while GPT-5.5 Instant continues to handle fast everyday conversations as the default model across the service.

The model picker now presents reasoning effort through labels such as Medium, High, Extra High, and Pro, which describe how much computation ChatGPT applies and, in the case of Pro, which GPT-5.6 model processes the request.

Free users remain outside GPT-5.6 Sol in ordinary chats, Plus becomes the lowest individual plan with direct Sol access, and the Pro tiers add the highest reasoning settings together with GPT-5.6 Sol Pro.

Business and Enterprise receive a similarly broad model range inside managed workspaces, although administrators, workspace settings, credit policies, and staged rollouts may determine which options an individual employee sees.

ChatGPT Team should no longer be treated as a separate current subscription, because OpenAI renamed it ChatGPT Business on August 29, 2025, while retaining the existing workspaces, contracts, features, pricing arrangements, usage limits, and security configuration at the time of the change.

·····

GPT-5.6 changes reasoning access without replacing the default chat model.

GPT-5.5 Instant remains the default selection for routine questions, short writing tasks, summaries, lightweight analysis, and other requests for which immediate response speed is prioritized over extended reasoning.

When automatic model routing is enabled on an eligible paid account, ChatGPT may move a sufficiently complex request from Instant to Medium, which causes GPT-5.6 Sol to process the prompt without requiring the user to select a reasoning model manually.

Users who prefer direct control can open the model picker and select Medium, High, Extra High, or Pro, although the available levels depend on the subscription and, within organizational workspaces, the models that administrators have enabled.

Medium, High, and Extra High all use GPT-5.6 Sol, while the distinction concerns reasoning effort rather than a change to another model family.

Pro uses GPT-5.6 Sol Pro, which OpenAI positions for difficult tasks and longer-running workflows that require the highest available standard-chat capability.

A higher effort setting generally allows the model to spend more time analysing intermediate steps, checking relationships between information, and developing a response before producing the final output, although it may also increase latency and consume a limited reasoning allowance more quickly.

........

Current ChatGPT model-picker labels and their underlying models.

Model-picker option	Underlying model	Intended processing level
Instant	GPT-5.5 Instant	Fast everyday responses
Medium	GPT-5.6 Sol	Standard reasoning
High	GPT-5.6 Sol	Extended reasoning
Extra High	GPT-5.6 Sol	Highest Sol reasoning effort
Pro	GPT-5.6 Sol Pro	Highest-capability option for difficult and longer-running work

·····

Free users do not receive GPT-5.6 Sol in standard ChatGPT conversations.

The Free plan continues to provide limited access to GPT-5.5 Instant, together with a variable set of lightweight reasoning options, file uploads, image generation, data analysis, search, memory, voice, and other features that operate under lower allowances than the paid plans.

Medium, High, Extra High, and Pro are unavailable in ordinary Free conversations because those selections require GPT-5.6 Sol or GPT-5.6 Sol Pro.

Free usage of GPT-5.5 Instant operates within a dynamic five-hour window, while the precise number of messages may vary according to system capacity, regional conditions, abuse-prevention measures, and individual usage patterns.

When the account reaches its current Instant allowance, ChatGPT may restrict access, route requests to a smaller fallback model, or display a reset time when that information is available.

Limited GPT-5.6 Terra access may still appear through ChatGPT Work and Codex on supported desktop experiences, which creates a narrower path into the GPT-5.6 family without providing Sol in the ordinary chat model picker.

Terra is designed as a lower-cost and faster member of the GPT-5.6 family, whereas Sol remains the flagship tier used for the Medium, High, and Extra High chat settings.

The presence of Terra in a desktop workflow should therefore not be described as full Free access to GPT-5.6, because the standard conversational product, the available reasoning effort, and the associated usage limits remain materially different from Plus and higher plans.

The Free plan also has a 27K total context window for Instant interactions, while its reasoning context is listed as variable, which limits how much conversation history, uploaded text, retrieved context, system instruction, memory, and output generation can fit into a single processing window.

·····

Plus becomes the entry subscription for GPT-5.6 Sol in ordinary chats.

ChatGPT Plus remains priced at $20 per month with monthly billing, and it is the lowest-priced individual subscription that places GPT-5.6 Sol directly inside the normal ChatGPT model picker.

Plus users receive Medium and High reasoning, both of which run on GPT-5.6 Sol, while Extra High and Pro remain unavailable.

Medium suits prompts that require deliberate analysis without the longest processing time, whereas High allocates additional reasoning effort to tasks involving multi-stage decisions, difficult coding, complex document interpretation, planning, or research synthesis.

Because both settings use the same underlying model, selecting High does not unlock a different training generation or a larger model family, although the additional inference effort may change the depth, latency, and consistency of the response.

Plus continues to use GPT-5.5 Instant as the everyday default, and OpenAI currently allows Plus and Go users to send up to 160 Instant messages within a three-hour period before conversations temporarily fall back to GPT-5.5 Instant mini.

That published Instant allowance should not be applied to manually selected GPT-5.6 Sol reasoning, because OpenAI maintains separate counters for automatic routing, manually selected reasoning, and certain higher-capability models.

OpenAI does not publish one permanent global number for Plus access to GPT-5.6 Sol, since limits may change with system conditions, rollout stages, account configuration, and ongoing capacity management.

The model picker or account interface may show when a reasoning allowance resets, while a conversation that reaches its GPT-5.6 limit may continue through GPT-5.4 Thinking mini or another available fallback.

Plus provides a 54K total context window for Instant and a 256K total context window for reasoning, although part of each window is reserved for system instructions, tools, memory, internal processing, retrieved information, and generated output.

OpenAI estimates that the Plus reasoning input capacity may correspond to approximately 320 pages of text under favourable formatting conditions, while dense tables, images, code, complex layouts, and lengthy instructions alter the practical amount that can be processed.

·····

Pro now separates capability from usage through two monthly prices.

ChatGPT Pro is available through a $100 monthly tier and a $200 monthly tier, both of which include the same core Pro capabilities while allocating different amounts of usage.

The $100 version provides approximately five times the Plus usage allowance, whereas the $200 version provides approximately twenty times the Plus allowance and remains the highest-usage individual subscription.

Both versions include Medium, High, and Extra High through GPT-5.6 Sol, together with the Pro option powered by GPT-5.6 Sol Pro.

A user does not need the $200 tier to obtain Sol Pro, because the difference between the two Pro prices concerns usage volume rather than access to a separate set of models.

Extra High represents the largest reasoning effort available through standard GPT-5.6 Sol, while Pro moves the request to GPT-5.6 Sol Pro for demanding workflows in which the highest available model quality outweighs response speed and allowance consumption.

The distinction becomes relevant for researchers, developers, analysts, and other frequent users whose workloads involve repeated long-context prompts, extensive deep research, agentic activity, large codebases, document production, or iterative reasoning across many sessions.

Both Pro subscriptions also expand Codex tasks, image generation, deep research, memory, context, projects, scheduled tasks, custom GPTs, and access to experimental features, although individual tools may retain their own rate limits or credit rules.

OpenAI describes Pro messaging and several included capabilities as unlimited, but that language remains subject to abuse guardrails, reasonable-use requirements, and model-specific allowances.

Automated extraction, account sharing, resale, credential sharing, or using a personal subscription to power a third-party service may trigger temporary restrictions even when the account has not encountered an ordinary usage counter.

Pro provides a 128K total context window for Instant and a 400K total context window for reasoning, with OpenAI estimating that the reasoning input may reach approximately 680 pages of text when formatting and reserved context leave sufficient capacity.

........

Current differences between the two ChatGPT Pro tiers.

Pro tier	Monthly price	Usage relative to Plus	GPT-5.6 Sol	Extra High	GPT-5.6 Sol Pro
Pro lower-usage tier	$100	Approximately 5×	Included	Included	Included
Pro highest-usage tier	$200	Approximately 20×	Included	Included	Included

·····

ChatGPT Team has become ChatGPT Business rather than remaining a separate plan.

OpenAI renamed ChatGPT Team to ChatGPT Business on August 29, 2025, which removed Team as a current independent subscription name.

Existing Team workspaces became Business workspaces without requiring customers to migrate their users, recreate projects, change workspace data, or sign replacement agreements solely because of the rename.

Features, pricing, usage limits, contracts, billing processes, and security arrangements remained unchanged at the moment of renaming, although OpenAI has continued to update Business pricing and seat structures through later product changes.

Invoices, product interfaces, account settings, administrative screens, and OpenAI communications now use the Business name, which may create confusion when older reviews, procurement records, tutorials, or internal documents still refer to Team.

A comparison that lists Team and Business as two current paid tiers would therefore count the same product twice and could imply differences that no longer exist.

References to Team remain relevant when describing historical contracts, older screenshots, legacy documentation, or the origin of a workspace, while current model access follows the Business rules.

Organizations that subscribed under the Team name should check their Business workspace model picker, because GPT-5.6 availability depends on the current plan, workspace configuration, administrator controls, and rollout status rather than the name shown on an older invoice.

·····

Business combines broad GPT-5.6 access with a managed workspace.

ChatGPT Business currently includes Medium, High, Extra High, and Pro in standard conversations, which places its published GPT-5.6 range alongside the Pro and Enterprise plans.

Medium, High, and Extra High use GPT-5.6 Sol, while Pro uses GPT-5.6 Sol Pro, although individual members may see fewer options when workspace administrators restrict model access or when a feature has not completed its rollout.

The Business model picker may also display Pro Standard and Pro Extended, which apply different reasoning effort to the Pro-tier model while remaining subject to workspace availability.

ChatGPT Business costs $25 per user per month with monthly billing or $20 per user per month when billed annually in most countries, and a workspace requires at least two standard ChatGPT seats.

Regional currency, taxes, local pricing, and purchasing arrangements may change the final amount shown at checkout.

A standard Business seat includes ChatGPT and baseline Codex access, together with GPTs, Projects, Apps, Company Knowledge, ChatGPT Agent, Deep Research, centralized billing, workspace administration, and controls for user roles and access.

Business workspaces receive virtually unlimited messages for eligible base models, although OpenAI applies policy guardrails, prevents account sharing and resale, and may impose temporary restrictions when usage resembles automated extraction or another prohibited pattern.

GPT-5.6 Sol has a documented 272K context window in Business, while GPT-5.6 Terra and Luna have 128K context windows within the workspace documentation.

Business customers may purchase workspace credits when included allowances are exhausted or when advanced features operate through flexible pricing, allowing an organization to continue selected work without changing every user to another subscription.

Workspace data is not used to train OpenAI models by default, which separates Business from consumer accounts whose users may need to manage training preferences through personal data controls.

........

Current Business pricing and workspace requirements.

Billing structure	Price in most countries	Minimum standard seats	GPT-5.6 standard-chat access
Monthly billing	$25 per user per month	2	Medium, High, Extra High, and Pro
Annual billing	$20 per user per month, billed annually	2	Medium, High, Extra High, and Pro

·····

Enterprise pairs GPT-5.6 access with contractual deployment controls.

ChatGPT Enterprise also includes Medium, High, Extra High, and Pro, giving eligible workspace members access to GPT-5.6 Sol and GPT-5.6 Sol Pro when administrators have enabled the corresponding options.

The model range does not create the central distinction between Business and Enterprise, because both plans currently support the broad set of GPT-5.6 reasoning levels in standard chats.

Enterprise instead addresses organizations that require negotiated contracts, invoicing, identity management, role-based access controls, domain verification, SCIM provisioning, analytics, audit capabilities, data residency, custom retention policies, administrative segmentation, support commitments, or legal and security terms that exceed the self-service Business plan.

Pricing is negotiated with OpenAI rather than displayed as a universal public per-user amount, because seat counts, deployment requirements, data controls, service terms, support, residency, and usage arrangements differ across customers.

Enterprise workspace owners may control model availability by role or group, which means that two employees within the same organization may receive different GPT-5.6 options even though the company has one Enterprise contract.

Early model access may also require an administrator to enable previews, while regulated environments may delay deployment until internal testing, security review, or compliance approval has been completed.

Flexible pricing allows Enterprise customers to draw from a shared credit pool for supported models and features, after which the organization may enable overages or purchase additional credits according to its contract.

Enterprise ChatGPT access remains separate from the OpenAI API, so an organization that needs programmatic GPT-5.6 calls must maintain an API organization, billing arrangement, permissions structure, and technical integration outside the ChatGPT subscription.

·····

The model picker now expresses reasoning effort rather than a list of model names.

Previous ChatGPT interfaces often required users to choose among several technical model names, whereas the GPT-5.6 interface translates much of that decision into speed and reasoning levels.

Instant prioritizes responsiveness through GPT-5.5 Instant, while Medium, High, and Extra High progressively increase the reasoning effort assigned to GPT-5.6 Sol.

Pro changes the underlying model to GPT-5.6 Sol Pro, which separates the highest-capability selection from the effort levels applied to standard Sol.

A user choosing between Medium and High is therefore deciding how long Sol should reason, while a user moving from High to Pro is selecting another GPT-5.6 model tier.

Automatic routing introduces another layer because ChatGPT may move a request from Instant to Medium when the prompt appears to require additional analysis.

OpenAI states that automatic switching does not consume the allowance reserved for manually selected reasoning, although flexible-pricing workspaces may still record credit consumption according to the model that processed the message.

A Business or Enterprise request that begins in Instant and automatically moves to Medium may therefore use GPT-5.6 Sol and consume the relevant workspace credits, even when it does not reduce the separate manual reasoning counter in the same manner.

The simplified labels make the interface easier to navigate, but they require users to distinguish between model identity, reasoning effort, message allowance, context size, and flexible-credit consumption.

·····

Usage limits remain separate across Instant, reasoning, Pro, and advanced tools.

The overall message allowance shown on a pricing page does not guarantee unlimited access to every model, because OpenAI maintains separate restrictions for selected reasoning models, Pro models, deep research, agent mode, Codex, image generation, voice, and other computationally intensive functions.

Free accounts receive dynamic GPT-5.5 Instant limits within a five-hour window, while Plus and Go currently receive up to 160 Instant messages every three hours before temporary fallback to GPT-5.5 Instant mini.

Manually selecting Medium, High, or Extra High activates GPT-5.6 Sol and draws from the plan’s reasoning allowance, which does not have one permanently published global message count for all users.

Selecting Pro activates GPT-5.6 Sol Pro and may draw from another allowance, whose availability depends on Pro, Business, or Enterprise access and the account’s current usage state.

When a reasoning allowance is exhausted, ChatGPT may display the reset time, continue through a smaller reasoning model, or invite the user to choose another available option.

Pro users receive five or twenty times the Plus usage allowance according to their selected monthly tier, although model-specific limits and abuse controls still apply.

Business and Enterprise customers may combine included access with flexible credits, which changes the practical ceiling from a fixed stop into a budget-controlled continuation for supported workloads.

Workspace owners can monitor or restrict spending, while employees may encounter a model limit even when the organization still has credits if the relevant administrative policy, role, or product setting prevents further use.

·····

Context windows differ substantially across individual and organizational plans.

A context window contains the prompt, relevant conversation history, uploaded material, retrieved information, system instructions, memory, tool output, internal processing space, and the model’s generated response.

The advertised total context therefore exceeds the amount available for a user’s raw document, because ChatGPT reserves part of the window for the surrounding system and workflow.

Free provides a 27K total Instant context window, while its reasoning context varies according to the available fallback model and current product configuration.

Plus provides 54K for Instant and 256K for reasoning, allowing longer reports, code files, transcripts, and multi-document tasks to remain within one model interaction.

Pro increases those figures to 128K for Instant and 400K for reasoning, which accommodates larger project histories and more extensive source material before compression, summarization, or omission becomes necessary.

Business documents a 272K context window for GPT-5.6 Sol and 128K for Terra and Luna, while Enterprise capacity may depend on the specific model, product surface, workspace configuration, and contract.

Page estimates should be treated as approximations because a page of plain prose consumes a different number of tokens from a page containing tables, code, citations, formulas, scanned text, or complex formatting.

A larger window also does not guarantee that every detail will receive equal attention, since retrieval quality, prompt structure, file parsing, tool selection, and the distribution of relevant information affect how the model uses the available context.

........

Published context figures for selected ChatGPT plans and GPT-5.6 environments.

Plan or workspace	Instant context	GPT-5.6 reasoning context
Free	27K	Variable
Plus	54K	256K
Pro	128K	400K
Business	Underlying Instant model may vary	272K for Sol; 128K for Terra and Luna
Enterprise	Depends on workspace and model configuration	Depends on contract, product, and model configuration

·····

ChatGPT Work and Codex create another layer of GPT-5.6 availability.

GPT-5.6 access in standard ChatGPT conversations does not describe every place where the model family appears, because ChatGPT Work and Codex use a separate availability structure.

Free and Go users may receive limited GPT-5.6 Terra access through supported Work and Codex desktop experiences even though they cannot select GPT-5.6 Sol in a normal conversation.

Plus, Pro, Business, and Enterprise users can choose among GPT-5.6 Sol, Terra, and Luna in Work and Codex, allowing them to balance capability, speed, and usage cost according to the task.

Sol serves as the flagship model for demanding professional work, Terra targets a lower-cost balance between speed and capability, and Luna prioritizes the fastest and least expensive processing within the family.

The max effort setting is available to users who have GPT-5.6 access in Work and Codex when the option is enabled, while ultra is available to Pro and Enterprise in ChatGPT Work and to Plus and higher plans in Codex.

A Plus subscriber may therefore receive a broader GPT-5.6 model selection inside Codex than in the ordinary chat picker, where the plan is limited to Medium and High through Sol.

Free access presents the opposite pattern, because limited Terra may appear in Work or Codex while every GPT-5.6 Sol reasoning level remains absent from standard conversations.

Product names must therefore accompany any statement about plan access, since “Free has GPT-5.6” and “Free has no GPT-5.6” are both incomplete without identifying whether the comparison concerns Chat, Work, Codex, or the API.

·····

Flexible pricing changes how Business and Enterprise usage is calculated.

Business and Enterprise workspaces may use a flexible-pricing structure through which included allowances are consumed first and additional activity draws from a shared pool of credits.

GPT-5.5 Instant is listed as unlimited for eligible base usage, while each GPT-5.6 Sol message consumes approximately ten credits under the current Business and Enterprise rate card.

Medium, High, and Extra High carry the same approximate ten-credit rate because they all use GPT-5.6 Sol, even though the selected effort changes the amount of reasoning applied to the request.

GPT-5.6 Sol Pro costs approximately fifty credits per message, creating a fivefold difference from Sol under the workspace credit schedule.

An Instant request that automatically routes to Medium uses GPT-5.6 Sol and may therefore consume approximately ten credits, despite beginning from the default fast setting.

Agent mode is listed at approximately thirty credits per message, deep research at fifty credits per task, image generation at five credits per generation, and voice at five credits per minute.

These figures apply to flexible pricing rather than creating an automatic surcharge for every Business or Enterprise interaction, because each workspace may receive included usage, negotiated allowances, credit pools, spending controls, or contract-specific arrangements.

........

Approximate Business and Enterprise flexible-credit rates.

Model or feature	Billing unit	Approximate credit rate
GPT-5.5 Instant	One message	Unlimited eligible base usage
GPT-5.6 Sol	One message	10 credits
GPT-5.6 Sol Pro	One message	50 credits
Agent mode	One message	30 credits
Deep research	One task	50 credits
Image generation	One generation	5 credits
Voice	One minute	5 credits

·····

ChatGPT subscriptions do not include GPT-5.6 API usage.

Access to GPT-5.6 through Plus, Pro, Business, or Enterprise does not include general API calls, because ChatGPT subscriptions and OpenAI API organizations use separate billing and access systems.

A developer who pays for ChatGPT Pro may use GPT-5.6 Sol Pro in the ChatGPT interface while still needing a funded API account to integrate GPT-5.6 into software, automation, internal tools, or customer-facing services.

Business and Enterprise customers face the same separation unless their organization has negotiated and configured an additional API arrangement.

Current API pricing lists GPT-5.6 Sol at $5 per million input tokens and $30 per million output tokens, Terra at $2.50 per million input tokens and $15 per million output tokens, and Luna at $1 per million input tokens and $6 per million output tokens.

GPT-5.6 cache writes are billed at 1.25 times the uncached input rate, while cache reads receive a ninety-percent discount from the ordinary input price.

API token prices should not be converted into estimated ChatGPT message counts, because the consumer and workspace subscriptions combine usage limits, product features, tools, model routing, context management, and capacity controls that do not correspond directly to API billing.

An application that needs guaranteed programmatic access, repeatable parameters, structured outputs, tool coordination, or production monitoring should use the API rather than attempting to automate a ChatGPT subscription.

·····

Plan selection depends on where GPT-5.6 enters the workflow.

Free remains suited to occasional conversations, search, lightweight file work, voice, basic image generation, and limited experimentation, while its standard chat experience does not include GPT-5.6 Sol.

Plus provides the lowest individual price for direct Sol reasoning, and its Medium and High settings cover many professional workflows involving writing, document analysis, planning, research, coding, and structured problem-solving.

The absence of Extra High and Sol Pro becomes relevant when a user repeatedly encounters tasks whose complexity, duration, or quality requirements exceed the Plus reasoning allowance.

The $100 Pro tier provides the complete individual GPT-5.6 range with five times the Plus usage, which may suit users who need Sol Pro but do not generate enough activity to justify the $200 tier.

The $200 Pro tier retains the same model choices while expanding usage to approximately twenty times the Plus allowance, making the additional payment primarily a capacity decision.

Business introduces a shared workspace, centralized administration, organization-level privacy defaults, Company Knowledge, user controls, spend management, and broad GPT-5.6 access for at least two standard seats.

Enterprise becomes relevant when deployment requires contractual security terms, identity provisioning, role-based permissions, data residency, audit capabilities, custom retention, invoicing, support commitments, or organization-wide governance.

Former Team customers already fall within the Business category, so they should evaluate current Business pricing, seat rules, model availability, and credit policies rather than searching for a separate Team upgrade path.

The live model picker remains the most precise account-level reference because staged rollouts, regional availability, workspace administration, temporary capacity limits, and product-specific access may produce differences between the general plan description and the options visible to an individual user.

·····

FOLLOW US FOR MORE.

·····

DATA STUDIOS

·····

[datastudios.org]

·····