Comparison
BlueAI, ChatGPT, Claude and Perplexity, limit for limit
Every column in one table and in the same units: how many messages fit in a burst, how many in a month, how many long messages and research sessions are included, and roughly how many API tokens all of that spends. Our numbers come from the module the service enforces with. Everyone else’s are quoted as published, with their own date underneath.
| Plan | Price | Messages, burst | Messages a month | Long messages | Deep research | API tokens |
|---|---|---|---|---|---|---|
| BlueAI | ||||||
| Free | $0 | 20 / 3 h | 600 | 12 / mo | 5 / mo | 1.2M / mo |
| Pro | $5/mo | 180 / 3 h | 10,500 | 360 / mo | 36 / mo | 36M / mo |
| Max x5 | $20/mo | 600 / 3 h | 42,000 | 1,500 / mo | 150 / mo | 150M / mo |
| Max x20 | $50/mo | 1,500 / 3 h | 108,000 | 3,600 / mo | 450 / mo | 360M / mo |
| ChatGPT | ||||||
| ChatGPT Free | $0 | 10 / 5 h | not published | none | 5 / mo, light | ≈ 1M |
| ChatGPT Go | $8/mo | 160 / 3 h | not published | 10 / 5 h | ≈ 10 / mo | ≈ 5M |
| ChatGPT Plus | $20/mo | 160 / 3 h | not published | 3,000 / week | 25 / mo | ≈ 12M |
| ChatGPT Pro | $200/mo | unlimited* | unlimited* | unlimited* | 250 / mo | ≈ 80M |
| Claude | ||||||
| Claude Free | $0 | limited, 5 h window | not published | none | none | ≈ 0.7M |
| Claude Pro | $20/mo | ≈ 45 / 5 h | not published | 200K context | included, weekly cap | ≈ 7M |
| Claude Max 5x | $100/mo | ≈ 225 / 5 h | not published | 200K context | included, weekly cap | ≈ 35M |
| Claude Max 20x | $200/mo | ≈ 900 / 5 h | not published | 200K context | included, weekly cap | ≈ 120M |
| Perplexity | ||||||
| Perplexity Free | $0 | 3 Pro searches / day | unlimited quick* | none | 5 / day | ≈ 0.8M |
| Perplexity Pro | $20/mo | 300+ Pro searches / day | unlimited quick* | Labs, 50 / mo | 500 / day | ≈ 15M |
| Perplexity Max | $200/mo | unlimited* | unlimited* | Labs, unlimited* | unlimited* | ≈ 50M |
ChatGPT limits as published by OpenAI, August 2026. OpenAI publishes burst windows and multipliers rather than monthly totals, so figures marked ≈ are our own estimates from those windows, not OpenAI's numbers. * fair-use limits still apply.
Claude limits as published by Anthropic, August 2026. Anthropic publishes rolling five-hour windows that move with demand, plus weekly caps per model, and no monthly message total at all, so every count marked ≈ is our own estimate rather than Anthropic's number.
Perplexity limits as published, August 2026. Perplexity counts Pro searches and Deep Research runs a day rather than messages a month, and leaves its quick model uncapped, so those cells keep Perplexity's unit instead of being converted into ours. * fair-use limits still apply.
Plan against plan
The head-to-heads below are against ChatGPT, the plan most people are leaving, and each one is matched with the ChatGPT tier at the same price or dearer — never a cheaper one.
Pro vs ChatGPT Go
For learning and using AI every week.
| Limit | Pro$5/mo | GPT Go$8/mo |
|---|---|---|
| Messages, burst | 180 / 3 h | 160 / 3 h |
| Messages a month | 10,500 | not published |
| Long messages | 360 / mo | 10 / 5 h |
| Deep research | 36 / mo | ≈ 10 / mo |
| API tokens | 36M / mo | ≈ 5M |
ChatGPT limits as published, August 2026; token figures are estimates. Full comparison
Max x5 vs ChatGPT Plus
For daily AI work, research, and stronger routines.
| Limit | Max x5$20/mo | GPT Plus$20/mo |
|---|---|---|
| Messages, burst | 600 / 3 h | 160 / 3 h |
| Messages a month | 42,000 | not published |
| Long messages | 1,500 / mo | 3,000 / week |
| Deep research | 150 / mo | 25 / mo |
| API tokens | 150M / mo | ≈ 12M |
ChatGPT limits as published, August 2026; token figures are estimates. Full comparison
Max x20 vs ChatGPT Pro
For heavy research and professional AI workflows.
| Limit | Max x20$50/mo | GPT Pro$200/mo |
|---|---|---|
| Messages, burst | 1,500 / 3 h | unlimited* |
| Messages a month | 108,000 | unlimited* |
| Long messages | 3,600 / mo | unlimited* |
| Deep research | 450 / mo | 250 / mo |
| API tokens | 360M / mo | ≈ 80M |
ChatGPT limits as published, August 2026; token figures are estimates. Full comparison
Questions about the table
Per dollar, yes, and the table above says exactly where. Pro at $5 allows 10,500 messages and 36M of tokens a month, against ChatGPT Go at $8, Claude Pro at $20 and Perplexity Pro at $20. Where a rival is genuinely unlimited — ChatGPT Pro at $200, Perplexity Max at $200 — the table says so instead of hiding it.
Because they are not. OpenAI states burst windows and multipliers, Anthropic states five-hour windows and weekly caps that move with demand, and neither publishes a monthly message total or a token budget. Rather than invent one and present it as theirs, the cell says what is missing. Figures marked ≈ are our own estimates from those published windows, at roughly 3,400 tokens for an ordinary turn.
Because that is the unit Perplexity sells: Pro searches and Deep Research runs a day, with the quick model left uncapped. Converting that into messages a month would produce a number Perplexity never published and we could not defend, so the cell keeps their unit. The token column is where the three sides become properly comparable.
A long message is about 100,000 tokens of context — a large document, a long thread, a codebase excerpt. A deep research session is roughly 500,000 tokens, because it reads many sources before it writes anything. Both are drawn from the same monthly budget as ordinary messages, which is why every plan lists all three.
Ours are read directly from the module the service enforces with, so they cannot drift from what your account actually gets. The competitor columns were last checked in August 2026, and each column carries its own date under the table. If you find one that has changed, tell us and it gets corrected the same day.