ChatGPT Plus vs Claude Pro — the same $20, a different plan.
ChatGPT Plus and Claude Pro both cost $20 a month. Only Claude Pro can be paid annually, at $17 a month for the year up front. Both include a coding agent — Codex on Plus, Claude Code on Pro — and both meter usage in five-hour windows. Plus is the broader plan (GPT-6 reasoning, deep research, ChatGPT Work); Pro is the one to pick if most of your use is code or long documents with Opus.
| Category | Winner | Margin |
|---|---|---|
| Monthly price · web, US | ·Same price | $20 against $20 |
| Annual billing · per month, paid yearly | BClaude Pro | Claude Pro $17 a month; OpenAI sells Plus monthly only |
| Coding agent · included in the plan | ·Both included | Codex on Plus, Claude Code on Pro — each counted against the plan's own windows |
| How usage is counted · limits | AChatGPT Plus | OpenAI publishes per-model ranges for each five-hour window; Anthropic publishes only multiples of its own tiers |
| Top model in the plan | ·Each lab's current family | GPT-6 lineup up to GPT-6.2 Astra on Plus; Opus, Sonnet and Haiku on Pro, with Fable only on paid credits |
| Research tools · deep research | AChatGPT Plus | Plus includes deep research and ChatGPT Work; Pro includes Research and Projects |
| The next step up · personal tier above | ·Same price | ChatGPT Pro 100 at $100, Claude Max at $100 |
| Team version · standard seat | ·Same price | Business $25 and Team $25 a seat, both two seats minimum |
If you want one plan for everything.
- GPT-6 reasoning models: GPT-6.2 Astra, GPT-6.2 Sol, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Astra and GPT-6 Luna
- Deep research and ChatGPT Work in the same plan
- Codex included, with published ranges per five-hour window for each model
- Free and Go below it if you want to step down later
- Monthly only — no year to commit to
If most of your use is code.
- Claude Code included, Anthropic's coding agent
- Opus, Sonnet and Haiku in one plan
- Annual billing at $17 a month
- Projects and Research for long documents
- Usage credits for the odd heavy day, at API rates
| Aspect | ChatGPT Plus | Claude Pro |
|---|---|---|
| Monthly priceindividual, web, before tax | $20/month Billed monthly | $20/month Billed monthly, or annually below |
| Annual billingper month | Not offered OpenAI does not support annual billing for Go, Plus or Pro | $17/month The year is paid up front B wins |
| Free plan below | Free and Go Go is $8 a month A wins | Free Sonnet and Haiku, no Claude Code |
| Plan abovenext personal tier | $100/month Pro 100; then Pro 200 and Pro 500 | $100/month Max; then Ultra at $200 |
| Team seatstandard seat, monthly | $25/seat $20 on annual billing, two seats minimum | $25/seat $20 on annual billing, two to 150 seats |
| Beyond the limit | Codex credits For Codex only; OpenAI publishes no price per credit | Usage credits Billed at API rates B wins |
| Capability | ChatGPT Plus | Claude Pro |
|---|---|---|
| Monthly price | $20 | $20 |
| Annual price | ✗ Monthly only | ✓ $17/mo |
| Flagship models | GPT-6.2 Astra, GPT-6.2 Sol, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Astra and GPT-6 Luna | Opus, Sonnet, Haiku |
| Top-tier model | GPT-6.2 Astra in Work and Codex; Pro reasoning needs Pro | Fable only on paid credits |
| Coding agent | ✓ Codex | ✓ Claude Code |
| Limit unit | Five-hour windows, per-model ranges published | Five-hour sessions plus a weekly limit |
| Codex / Claude Code allowance | 15–160 local messages per window on GPT-6.1 Sol | Shared with chat, no count published |
| Deep research | ✓ | ✓ Research |
| Workspace agent | ✓ ChatGPT Work | Projects |
| Pay past the limit | Codex credits, price not published | ✓ Usage credits at API rates |
| Cheaper tier below | ✓ Go $8 | ✗ Free only |
| Team seat | $25 / $20 annual | $25 / $20 annual |
| API included | ✗ Billed separately | ✗ Billed separately |
ChatGPT Plus
The wider plan: reasoning, research, Work and Codex for one monthly price.
Strengths
- GPT-6 reasoning up to GPT-6.2 Astra
- Deep research and ChatGPT Work in the same plan
- Codex with published per-model ranges for each five-hour window
- A cheaper tier below: Go at $8
Weaknesses
- No annual discount: the monthly price is the only price
- Past the limit, credits of unknown price: Codex credits can be bought, but OpenAI does not print what a credit costs
Best for
- Mixed daily use
- Research-heavy work
- Codex users
- People who want to cancel any month
Claude Pro
The coding plan of the two: Claude Code included, and the only one with an annual price.
Strengths
- Claude Code in the plan
- Opus, Sonnet and Haiku together
- $17 a month on annual billing
- Usage credits at API rates when you run out
Weaknesses
- No published counts: Anthropic sizes plans as multiples, not messages
- Fable is extra: only on paid usage credits, inside the plan from Max up
Best for
- Developers using Claude Code
- Long documents with Opus
- A year's commitment for a lower price
- Writing and analysis
A bit of everything, every day
You write, search, summarise and ask the odd coding question, a few dozen times a day.
Reasoning: Plus covers the widest set of tools for the money — deep research, Work, reasoning models — and you can stop any month. Both plans handle this load inside their limits.
An agent in your repository
You want a coding agent working in your codebase for a few focused hours a day.
Reasoning: Both include one. Claude Pro is the more common pick for Claude Code users and is cheaper on annual billing; Plus is the pick if your tools are built around Codex.
You will keep it for a year
You already know you will pay for the plan all year.
Reasoning: Only Claude Pro has an annual price: $17 a month instead of $20. ChatGPT Plus costs $20 every month either way.
Research reports
You need long, sourced reports on a topic several times a week.
Reasoning: Plus includes deep research alongside the reasoning models; Claude's Research does similar work, but Plus is built around it.
A team of three
Three people want the same assistant on one bill, with an admin console.
Reasoning: ChatGPT Business costs $25 a seat a month, or $20 on annual billing; Claude Team costs $25 or $20. Both need at least two seats. A Business seat gets the same usage table as Plus; a Claude Team seat gets 1.25 times Pro's usage per session and includes Claude Code. Claude Team is billed by card only.
You hit the limit every afternoon
Your heaviest work comes in one block, and the plan stops you before you are done.
Reasoning: Both plans count usage per five hours, not per month. OpenAI publishes ranges for Codex and Work on Plus: roughly 15–160 local messages on GPT-6.1 Sol in each window and 350–3,000 on GPT-6 Luna, so a lighter model stretches the same window. Anthropic publishes no counts; Claude Pro gives at least five times Free's usage per five-hour session, with a weekly cap on top. The next step costs the same on both sides: Pro 100 at $100, Max at $100.
Frequently asked.
Common questions about this comparison, with sources where they matter.