Claude vs ChatGPT: limits compared
Same prices, same five-hour window, but not the same rules. Here are both offers side by side, for chat and for code, according to Anthropic and OpenAI.
Plans
| Claude (Anthropic) | ChatGPT (OpenAI) | |
|---|---|---|
| Entry plan | Pro · $20 a month | Plus · $20 a month |
| 5x tier | Max 5x · $100 | Pro 5x · $100 |
| 20x tier | Max 20x · $200 | Pro 20x · $200, closed to new sign-ups since September 10, 2026 |
| Cheaper | Free | Free, and Go at $8 |
Prices in US dollars, before tax, published on September 29, 2026. Prices vary by country.
Limits
| Claude | ChatGPT and Codex | |
|---|---|---|
| Short window | Rolling 5-hour session | Estimates per 5-hour period |
| Week | Weekly limit, fixed day and time per account | “Weekly limits may also apply” |
| What’s shared | claude.ai, Claude Code and the desktop app draw on the same limit | Codex local messages and cloud tasks share the plan’s allowance |
| What’s published | A ratio: Max gives 5 or 20 times Pro per session | Ranges: for example 15 to 150 GPT-6 Sol messages per 5 hours on Plus |
Seeing your usage, and at the cap
| Claude | ChatGPT and Codex | |
|---|---|---|
| See your usage | Settings > Usage, or /usage in Claude Code | The Codex usage page, or /status in the CLI |
| At the cap | Wait, a higher plan, or usage credits | Wait, a higher plan, or credits valid for 12 months |
| Most expensive models | Fable: up to 50% of the weekly limit on Max, usage credits on Pro | Message ranges vary by model: Astra, GPT-6 Sol, GPT-6 Luna |
Which one?
At the same prices, the difference is in the work. For code, compare Claude Code and Codex on your own tasks: both limits are measured in real working hours, not messages. Many people keep a plan on each side and spread work according to what’s left.
Frequently asked questions
Do Claude Max and ChatGPT Pro cost the same?
Yes in dollars: $100 for the 5x tier and $200 for the 20x tier. OpenAI paused new sign-ups to Pro 20x on September 10, 2026.
Which gives more usage?
There’s no telling in numbers: Anthropic publishes a ratio between its plans, OpenAI message ranges that depend on task size. The only reliable comparison is your own usage.
Can I track both in one place?
Yes: JaugeAI shows Claude and Codex limits side by side.