ChatGPT Plus and Claude Pro both cost $20/month. The sticker price is identical, so the decision comes down to what each tier unlocks and how fast you hit the wall. ChatGPT Plus caps GPT-5.5 at 160 messages every 3 hours; Claude Pro has no per-message count and instead enforces a 5-hour session limit plus a weekly cap. This page covers every tier, every limit, and the June 2026 model lineup. For the capability question (which model is smarter at coding, writing, and reasoning) see Claude vs ChatGPT: benchmarks and pros/cons.
The 30-Second Answer
If you want the cheapest paid AI chatbot, ChatGPT Go at $8/month undercuts everything Claude offers, since Claude has no tier between Free and the $20 Pro plan. If you want the cheapest $20 tier, Claude Pro wins on annual billing: $17/month billed yearly ($204/year) versus $20/month flat for ChatGPT Plus.
If you care about how often you get cut off, the two plans throttle differently. ChatGPT Plus is a hard count: 160 GPT-5.5 messages every 3 hours, then a forced drop to the mini model. Claude Pro is a rolling budget: a 5-hour session limit plus a weekly cap across all models. Heavy short bursts favor Claude Pro; predictable message-by-message usage favors ChatGPT Plus.
If privacy is the deciding factor, Claude is the safer default. ChatGPT trains on your chats out of the box. Claude makes model-improvement training a user choice, and with it off your data falls under a 30-day retention window instead of a multi-year training pipeline.
Same price, different ceiling
$20 buys GPT-5.5 with a 160-per-3-hour cap on ChatGPT Plus, or Claude Fable 5 with a 5-hour session limit plus a weekly all-model cap on Claude Pro. Neither plan publishes a single number that makes one strictly more generous; the right pick depends on whether your usage is bursty (Claude) or steady (ChatGPT).
Every Tier, Side by Side
ChatGPT runs five consumer-facing price points; Claude runs four. The biggest structural difference is the bottom of the ladder: ChatGPT has an $8/month Go tier and an ad-supported Free tier, while Claude jumps straight from a fully free plan to $20 Pro.
| Tier | ChatGPT | Claude |
|---|---|---|
| Free | $0/mo, ads in the US | $0/mo, no ads |
| Entry paid | Go: $8/mo, ad-supported, 98 countries | No equivalent tier |
| Standard ($20) | Plus: $20/mo | Pro: $20/mo, or $17/mo annual |
| Mid premium | Pro: $100/mo, 5x Plus usage | Max 5x: from $100/mo, 5x Pro usage |
| Top premium | Pro Max: $200/mo, 20x Plus usage | Max 20x: 20x Pro usage, monthly only |
The $100 and $200 tiers line up on usage multipliers (5x and 20x), so at the premium end the choice is about features, not headline price. ChatGPT Pro Max adds GPT-5 Pro and unlimited audio and video mode. Claude Max adds priority traffic and early feature access, with no annual discount on either Max tier.
Free: ChatGPT vs Claude
Both companies let you use the product at $0. The difference is the trade-off. ChatGPT Free is ad-supported in the US and meters the flagship model tightly: up to 10 GPT-5.5 messages every 5 hours. Claude Free carries no ads and is broader on features, with no published per-message flagship cap of that kind.
| Feature | ChatGPT Free | Claude Free |
|---|---|---|
| Price | $0/mo, ads in the US | $0/mo, no ads |
| Flagship model access | Up to 10 GPT-5.5 messages / 5 hours | Multiple models, availability varies |
| Platforms | Web, iOS, Android, Desktop | Web, iOS, Android, Desktop |
| Web search | Yes | Yes |
| Memory | Yes | Yes |
| File creation / connectors | Limited | File creation, extensions, connectors |
For occasional questions, both free plans are enough. If you want the newest OpenAI model in short bursts without paying, ChatGPT Free gets you GPT-5.5 (10 per 5 hours). If you want an ad-free experience with file creation, code generation, and connectors at $0, Claude Free is the broader plan.
$20/mo: ChatGPT Plus vs Claude Pro
This is the tier most people actually compare. The price is the same monthly, but Claude Pro is cheaper annually at $17/month ($204/year), and the two plans bundle different things.
| Feature | ChatGPT Plus | Claude Pro |
|---|---|---|
| Monthly price | $20/mo | $20/mo |
| Annual price | $20/mo (no annual discount) | $17/mo ($204/year) |
| Flagship model | GPT-5.5 | Claude Fable 5 (and Opus, Sonnet, Haiku) |
| Coding agent | Codex Agent | Claude Code |
| Agent / automation | ChatGPT Agent | Claude Cowork |
| Research | Deep Research | Research |
| Video / image | Sora 1 preview (limited video) | No native video or image generation |
| Other | Canvas, Advanced Data Analysis, Vision, custom GPTs | Unlimited projects, Microsoft 365 / Outlook |
ChatGPT Plus is the broader media product: it bundles a Sora 1 video preview, Canvas, Advanced Data Analysis, Vision, and custom GPTs. Claude Pro is the deeper work product: Claude Code, Claude Cowork, unlimited projects, and Microsoft 365 / Outlook integration. If you generate or edit video and images inside the chat, Plus is the only one that does it. If you run coding and document workflows, Pro bundles more of that and costs less per year.
Message Caps and Usage Limits
This is where the two plans diverge most, and where most cancellations come from. ChatGPT publishes message counts. Claude publishes time-window and weekly budgets. They are not directly comparable, so here is each in its own terms.
ChatGPT message limits
- Free: up to 10 GPT-5.5 messages every 5 hours.
- Plus and Go: up to 160 GPT-5.5 messages every 3 hours; after the limit, chats switch to the mini version of the model until reset.
- GPT-5.5 Thinking (Plus or Business): manually selectable, up to 3,000 messages per week.
Claude usage limits
- Pro: a 5-hour session limit, plus a weekly usage limit across all models that resets seven days after the session starts. Both are viewable in Settings, Usage.
- Max: two weekly limits, one across all models and a second for Sonnet models only, both resetting seven days after the session starts.
Why the limits feel different
ChatGPT counts individual messages, so a long back-and-forth chews through your 160-per-3-hour budget one turn at a time. Claude meters compute over a rolling window, so a few heavy long-context turns can cost more of your budget than many short ones. Bursty, heavy sessions tend to favor Claude Pro; steady message-by-message use tends to favor ChatGPT Plus.
Teams and Business Pricing
For multi-seat deployments, the entry prices nearly match, with ChatGPT requiring at least two seats and Claude requiring at least five.
| Plan | ChatGPT | Claude |
|---|---|---|
| Standard seat (annual) | Business: $20/user/mo, min 2 users | Team Standard: $20/user/mo, 5-150 people |
| Standard seat (monthly) | Business: $25/user/mo | Team Standard: $25/user/mo |
| Premium seat | See ChatGPT Pro tiers | Team Premium: $100/mo annual, $125/mo monthly |
| Enterprise | SAML SSO, MFA, admin, compliance, connectors | From $20/seat + usage; SCIM, audit logs, HIPAA-ready |
ChatGPT Business is $20/user/month on annual billing or $25/user/month monthly (minimum 2 users; the price dropped to this on April 2, 2026) and adds SAML SSO, MFA, admin controls, compliance features, and connectors to Google Drive, SharePoint, GitHub, and Notion. Claude Team runs 5 to 150 people with a $20/$25 standard seat and a $100/$125 premium seat, and adds SSO, enterprise search, admin controls, and desktop deployment. Claude Enterprise lists $20/seat with usage cost scaling by model and task, plus SCIM, role-based access, audit logs, a Compliance API, custom retention, network controls, HIPAA-ready handling, and Claude Security (beta).
June 2026 Model Lineup
The flagship models behind each paid tier as of June 2026:
| Provider | Latest model | Released | Built for |
|---|---|---|---|
| ChatGPT | GPT-5.5 | API April 24, 2026 | General + agentic, with Pro variant |
| Claude | Fable 5 | June 9, 2026 | Demanding reasoning, long-horizon agentic |
GPT-5.5 is the latest GPT model live in ChatGPT; GPT-5.5 and GPT-5.5 Pro reached the API on April 24, 2026. On the benchmarks OpenAI publishes, GPT-5.5 scores 58.6% on SWE-Bench Pro (real-world GitHub issue resolution) and 82.7% on Terminal-Bench 2.0 (complex command-line workflows).
Claude Fable 5 became available June 9, 2026, and is Anthropic's most capable widely released model, built for demanding reasoning and long-horizon agentic work and state-of-the-art on nearly all tested capability benchmarks. Fable 5 is the same underlying model as Mythos 5, with added cybersecurity and biology safeguards; queries in those domains are automatically routed to Opus 4.8. For reference, the prior flagship Opus 4.5 scored 80.9% on SWE-bench Verified and 87.0% on GPQA Diamond.
GPT-5.5: 58.6% SWE-Bench Pro
OpenAI's latest model also reaches 82.7% on Terminal-Bench 2.0 for command-line workflows.
Claude Fable 5: June 9, 2026
Anthropic's most capable widely released model, state-of-the-art on nearly all tested capability benchmarks.
Both ship at $20/mo
GPT-5.5 on ChatGPT Plus, Claude Fable 5 on Claude Pro, with different message and usage caps.
Data Training Defaults
The default behavior differs, and it matters if you paste anything sensitive into a consumer chat.
On ChatGPT, the "Improve the model for everyone" toggle is on by default for consumer accounts. You can opt out in Settings, Data Controls, by switching it off, after which new conversations are not used to train models. The setting applies to the whole account regardless of device and can be changed anytime with no restrictions; conversations still appear in chat history but are not used for training.
On Claude consumer plans (Free, Pro, Max), model-improvement training is a user choice. If the model-improvement setting is off, Anthropic does not use previous or new chats and coding sessions for future training. If you allow chats to improve Claude, data may be retained de-identified for up to 5 years in training pipelines; if you do not, the standard 30-day data retention period applies.
| Setting | ChatGPT | Claude |
|---|---|---|
| Trains on chats by default | Yes (consumer) | User choice |
| How to opt out | Settings, Data Controls, toggle off | Turn off model-improvement setting |
| Retention if opted out | Chats stay in history, not used for training | 30-day retention |
| Retention if opted in | Used to improve the model | De-identified up to 5 years |
Frequently Asked Questions
Is ChatGPT Plus or Claude Pro cheaper?
Both are $20/month. Claude Pro is cheaper annually at $17/month ($204/year); ChatGPT Plus has no annual discount. Below Plus, ChatGPT also offers Go at $8/month (ad-supported, 98 countries). Claude has no tier between Free and the $20 Pro plan, so for the absolute cheapest paid AI chatbot, ChatGPT Go wins.
What is the message limit on ChatGPT Plus vs Claude Pro?
ChatGPT Plus allows up to 160 GPT-5.5 messages every 3 hours, then drops to the mini model until reset. GPT-5.5 Thinking on Plus is capped at 3,000 messages per week. Claude Pro does not publish a per-message count: it uses a 5-hour session limit plus a weekly usage limit across all models, both viewable in Settings, Usage.
What does ChatGPT Plus include for $20 a month?
GPT-5.5, the ChatGPT Agent, Codex Agent, a Sora 1 preview for limited video, Canvas, Advanced Data Analysis, Deep Research, Vision, and the ability to create and share custom GPTs.
What does Claude Pro include for $20 a month?
Claude Code, Claude Cowork, unlimited projects, Research, access to multiple models (Fable, Opus, Sonnet, Haiku, with availability by tier), and Microsoft 365 and Outlook integration. Annual billing drops it to $17/month.
ChatGPT Pro vs Claude Max: which premium plan is better?
ChatGPT Pro is $100/month (5x Plus, elevated Codex limits, 125 Deep Research queries per month); Pro Max is $200/month (20x Plus, GPT-5 Pro, unlimited audio and video). Claude Max 5x starts at $100/month (5x Pro) and Max 20x gives 20x Pro usage, monthly-only with no annual discount. The usage multipliers match; pick on features (video on ChatGPT, priority traffic and early access on Claude).
Is the free version of ChatGPT or Claude better?
ChatGPT Free ($0, ads in the US) gives up to 10 GPT-5.5 messages every 5 hours. Claude Free ($0, no ads) is broader: web, iOS, Android, Desktop chat, code generation, web search, memory, file creation, extensions, and connectors. ChatGPT Free unlocks the flagship in short bursts; Claude Free is broader on features.
Does ChatGPT or Claude train on my chats by default?
ChatGPT trains by default (the "Improve the model for everyone" toggle is on); opt out in Settings, Data Controls. Claude makes it a user choice; with it off, the standard 30-day retention applies, and with it on, data may be retained de-identified up to 5 years.
Related Resources
Building on Claude or GPT? Route Per Request.
If you are paying for a consumer plan to test models, the API tells you more. Morph Router classifies prompt difficulty and picks the right model tier across Anthropic and OpenAI automatically. $0.001 per request, about 430ms of added latency.