ChatGPT vs Claude Pricing: Plans and API Costs

Choose the subscription that completes your daily work; the plan table shows fees and separate access-verification dates. API usage is a different budget.

ChatGPT and Claude personal subscription fees

Plan	Recorded price	Tier	Published limits and access
ChatGPT Plus	$20/mo Price: Source verified: 2026-11-12 01:07:22 UTC. Official source	pro	Current pricing matrix lists GPT-6.1 Sol, GPT-6 Astra, GPT-6 Luna, and GPT-5.6 Terra. Model and tool limits vary by plan, rollout, and account; API billing is separate. Access and limits: Source verified: 2026-11-12 01:09:45 UTC. Official source
Claude Pro	$20/mo Price: Source verified: 2026-11-12 01:16:49 UTC. Official source	pro	At least 5x Free per five-hour session, with weekly limits. Outside the limited-time launch offer, Fable requires separate usage credits. Claude Duet research preview requires waitlist access, initially for US Pro, Max, and Ultra subscribers; shares existing plan limits, including background checks and tasks. No separate Duet allowance. Fable 5.2 launch offer: November 9–16, 2026, ending 11:59 pm PT on November 16. All Fable models temporarily included up to 50% of shared weekly limits; after the window or allowance, usage credits apply if enabled. One Fable reset for paid billing and seat membership before November 9 at 10 am PT; Team self-serve trials count as paid. Reset expires with the window; unused allowance does not roll over. Access and limits: Source verified: 2026-11-12 01:16:49 UTC. Official source
ChatGPT Pro ($100)	$100/mo Price: Source verified: 2026-11-12 01:07:21 UTC. Official source	power	Any unlimited Pro access is subject to abuse guardrails. Work and Codex allowance doubled automatically; weekly GPT-6 Pro message limits are unchanged. Pro 100 does not include Ultrafast; buying credits does not unlock it. Allowances differ by tier. Includes $100/month in claimable API credits. Claim after seven paid days by linking an API organization owned under the same email. First grant is prorated; unused credits expire at billing-cycle end without rollover. Credits cover eligible API Platform usage, not Codex via API keys or extra ChatGPT/signed-in Codex limits. API usage uses purchased credits after free credits; requests fail if no balance remains. Free trials and complimentary subscriptions do not qualify. The benefit may change or end; API terms and rate limits apply. Access and limits: Source verified: 2026-11-12 01:07:21 UTC. Official source
ChatGPT Pro ($200)	$200/mo Price: Source verified: 2026-11-12 01:07:21 UTC. Official source	power	Any unlimited Pro access is subject to abuse guardrails. The previously announced Pro 200 reduction is canceled; all subscribers now receive the same higher Work and Codex allowance. The lower allowance doubled; existing higher allowances remain unchanged. The lower weekly GPT-6 Pro message limit now matches the higher limit; existing higher weekly limits remain unchanged. Pro 200 does not include Ultrafast; buying credits does not unlock it. Allowances differ by tier. Includes $200/month in claimable API credits. Claim after seven paid days by linking an API organization owned under the same email. First grant is prorated; unused credits expire at billing-cycle end without rollover. Credits cover eligible API Platform usage, not Codex via API keys or extra ChatGPT/signed-in Codex limits. API usage uses purchased credits after free credits; requests fail if no balance remains. Free trials and complimentary subscriptions do not qualify. The benefit may change or end; API terms and rate limits apply. Access and limits: Source verified: 2026-11-12 01:07:21 UTC. Official source
Claude Max	$100/mo Price: Source verified: 2026-11-12 01:16:49 UTC. Official source	power	5x Pro usage every five hours AND every week. Fable uses up to half the shared weekly allowance, not an additional pool. Continue beyond included limits with separately paid usage credits. Includes $100/month in claimable Claude Platform API credits. $100/month Claude Platform API credits. Claim after seven days on an active eligible plan by linking one Console organization. Credits refresh monthly, expire without rollover and cover direct Claude Platform API usage, not interactive Claude Code, app extra usage or third-party clouds. API requests stop without other credits when exhausted; API billing is separate. Claude Duet research preview requires waitlist access, initially for US Pro, Max, and Ultra subscribers; shares existing plan limits, including background checks and tasks. No separate Duet allowance. Fable 5.2 launch: 1–3 Fable resets (under 6 months: 1; 6–12 months: 2; over 12 months: 3), based on first paid bill. Requires paid billing before November 9, 2026, 10 am PT and, for Team, seat membership before that cutoff. Resets expire March 8, 2027 at 11:59 pm PT; restore Fable weekly usage and any five-hour window, never raise limits. Canceling or downgrading forfeits unused resets. Access and limits: Source verified: 2026-11-12 01:16:49 UTC. Official source
Claude Ultra	$200/mo Price: Source verified: 2026-11-12 01:16:49 UTC. Official source	max	10x Pro usage every week. No five-hour cap for verified accounts; unverified accounts have a temporary limit, removed automatically after account verification. Fable uses up to half the shared weekly allowance, not an additional pool. Continue beyond included limits with separately paid usage credits. Includes $200/month in claimable Claude Platform API credits. $200/month Claude Platform API credits. Claim after seven days on an active eligible plan by linking one Console organization. Credits refresh monthly, expire without rollover and cover direct Claude Platform API usage, not interactive Claude Code, app extra usage or third-party clouds. API requests stop without other credits when exhausted; API billing is separate. Claude Duet research preview requires waitlist access, initially for US Pro, Max, and Ultra subscribers; shares existing plan limits, including background checks and tasks. No separate Duet allowance. Fable 5.2 launch: 1–3 Fable resets (under 6 months: 1; 6–12 months: 2; over 12 months: 3), based on first paid bill. Requires paid billing before November 9, 2026, 10 am PT and, for Team, seat membership before that cutoff. Resets expire March 8, 2027 at 11:59 pm PT; restore Fable weekly usage and any five-hour window, never raise limits. Canceling or downgrading forfeits unused resets. Access and limits: Source verified: 2026-11-12 01:16:49 UTC. Official source
ChatGPT Pro 500	$500/mo Price: Source verified: 2026-11-12 01:07:21 UTC. Official source	max	Any unlimited Pro access is subject to abuse guardrails. Work and Codex allowance doubled automatically; weekly GPT-6 Pro message limits are unchanged. Highest included usage among Pro tiers. GPT-6 Astra and GPT-6.1 Sol Ultrafast consume included allowance at 8x Standard and purchased credits at 6x Standard; these are usage-accounting multipliers, not speed guarantees. Includes $500/month in claimable API credits. Claim after seven paid days by linking an API organization owned under the same email. First grant is prorated; unused credits expire at billing-cycle end without rollover. Credits cover eligible API Platform usage, not Codex via API keys or extra ChatGPT/signed-in Codex limits. API usage uses purchased credits after free credits; requests fail if no balance remains. Free trials and complimentary subscriptions do not qualify. The benefit may change or end; API terms and rate limits apply. Access and limits: Source verified: 2026-11-12 01:07:21 UTC. Official source

Pricing data: How we verify →

Price and access verification are shown separately for each plan. Prices and plan notes are rendered at build time from the maintained subscription dataset.

By AI Pricing Guru Editorial Team

Updated November 9, 2026

Quick verdict

Choose the subscription that completes your daily work; the plan table shows fees and separate access-verification dates. API usage is a different budget.
Compare Plus with Pro first, then the heavier personal tiers only if your trial repeatedly reaches a relevant limit.
The API chart compares exact catalog models, not the models included in either subscription. Evaluate accepted results before choosing a provider.

Compare recorded API token rates separately

USD per 1M tokens. Input and output rates are charted separately; inclusion is not an availability or subscription-access guarantee.

InputOutput

Oldest price evidence in this chart: Source verified: 2026-11-12 01:07:03 UTC.

Estimate a separate API workload

Assumes 75% input tokens and 25% output tokens using current per-million rates.

Monthly tokens

GPT-6 Luna

openai

$2.00

Input share: $0.75
Output share: $1.25

Claude Haiku 5.5

anthropic

$2.00

Input share: $0.75
Output share: $1.25

GPT-6.1 Sol

openai

$40.00

Input share: $15.00
Output share: $25.00

Claude Sonnet 5.5

anthropic

$40.00

Input share: $15.00
Output share: $25.00

Claude Opus 5.5

anthropic

$80.00

Input share: $30.00
Output share: $50.00

Claude Fable 5.2

anthropic

$160.00

Input share: $60.00
Output share: $100.00

Separately billed API models — not plan entitlements

Model	Provider	Input / 1M	Cached / 1M	Output / 1M
GPT-6 Luna Price: Source verified: 2026-11-12 01:07:03 UTC. Official source	openai	$0.1	$0.01	$0.5
Claude Haiku 5.5 Price: Source verified: 2026-11-12 01:07:03 UTC. Official source	anthropic	$0.1	$0.01	$0.5
GPT-6.1 Sol Price: Source verified: 2026-11-12 01:07:03 UTC. Official source	openai	$2.00	$0.1	$10.00
Claude Sonnet 5.5 Price: Source verified: 2026-11-12 01:07:03 UTC. Official source	anthropic	$2.00	$0.1	$10.00
Claude Opus 5.5 Price: Source verified: 2026-11-12 01:07:03 UTC. Official source	anthropic	$4.00	$0.2	$20.00
Claude Fable 5.2 Price: Source verified: 2026-11-12 01:07:03 UTC. Official source	anthropic	$8.00	$0.2	$40.00

Pricing data: How we verify →

Built from pricing.json at publish time. Each row shows its own source-verification date; publication time is not a source check.

Affiliate disclosure: the optional Novita link below is sponsored. We may earn a commission at no extra cost to you.

Which subscription should you test?

Start with the same recurring tasks in both products: a document, a coding problem, a research question or a multimodal workflow you actually need. Compare successful output, review time and interruptions, not a provider-wide quality claim.

Your situation	What to compare	Buying check
Individual daily assistant	ChatGPT Plus and Claude Pro	Required tools, current access and actual usage headroom
Repeatedly blocked personal work	The heavier personal tiers above	Whether extra capacity removes your specific bottleneck
Coding workflows	Your required Codex or Claude Code integration	Product-specific access, billing and reset terms
Team procurement	Dedicated business/team plans	Seats, administration, data controls and minimum commitments
Software integration	The developer APIs	Metered usage, permissions and spending controls

These are evaluation criteria, not promises that a named feature is included in every listed tier. Use the ChatGPT subscription guide and Claude subscription guide for maintained plan details and official sources.

For current plans, compare Claude Max and Claude Ultra. Ultra replaces Max 20x automatically at the same subscription price; its 10× Pro weekly allowance is not a 20× weekly promise. Only verified Ultra accounts have no five-hour cap; unverified accounts retain a temporary cap. Max has 5× Pro usage every five hours and every week. See the Ultra launch report.

Read price and access separately

The subscription table is generated from the maintained catalog. Each row displays its own price verification and access verification. An article update, an API model release or a provider’s old announcement does not refresh either clock.

A successor API model must not be assumed to appear in an existing chat plan. Check the current official plan matrix, rollout and account eligibility before buying for one model or tool. Treat a record marked overdue as a previous observation, not a current quote.

The tracked Claude Max and Ultra rows include claimable Claude Platform API credits. Check redemption and billing conditions separately; this does not make all API usage unmetered.

The table covers tracked personal subscriptions, not a complete team procurement comparison. Check billing period, tax, currency, cancellation and any additional usage separately. Do not multiply a personal fee into a team quote without checking the actual team plan.

API costs are a different comparison

The chart, API table and mini-calculator use exact model identifiers from the pricing dataset. They do not establish subscription entitlement, interchangeable capabilities or equal task quality. Older migration-only models have been removed from this article’s default shortlist.

For routine extraction or classification, test the smaller candidates first. For coding, review and difficult tasks, test the larger candidates against a representative acceptance set. Escalate when the cheaper route fails your checks, not simply because a premium option exists.

Estimate uncached input, cached input, cache creation, output, retries and tool calls. Request-level context tiers and service modes can change the rate; monthly totals alone do not establish which requests cross a threshold.

Use OpenAI pricing, Anthropic pricing and the full token calculator. Our API cost calculation guide explains the arithmetic. Keep subscriptions and API budgets separate unless the exact current product terms explicitly establish otherwise.

Choose on accepted work

Run a short pilot before annual billing. Record accepted results, time spent fixing output, recurring limits and total paid usage. Choose the product that removes a real workflow cost; a larger allowance is poor value if you do not use it.

For a third API control, test a hosted open-model endpoint on Novita AI. Verify its exact model, usage terms and price independently. This is an API alternative, not a ChatGPT or Claude subscription replacement.

FAQ

Is ChatGPT cheaper than Claude?

Compare the relevant subscription rows and your actual usage requirements. API token rates answer a different question and do not establish the cheaper chat plan for your workflow.

Does a chat subscription include API credits?

Claude Max and Ultra list claimable API credits; see their current plan rows and redemption conditions. Do not transfer that benefit to other subscriptions or treat it as unlimited developer usage.

Does a new API model become available in my plan?

Not automatically. Confirm the model on the current plan-specific source and in your eligible account.

Which should I use for coding?

Trial your required integration on the same repository tasks. Compare accepted changes, review time, limits and total billing rather than choosing from a provider label.