AI message limits: how many messages each plan really gets

How Quad's limit works

Each plan comes with a monthly AI budget, shown in the app as credits. Every message spends part of that budget based on the model's price and the length of the message. There are no separate caps per model, no rolling five-hour windows and no weekly reset. What you get is one pool for the month that every model draws from. Which models you can pick depends on your plan. Claude Sonnet 5.5 starts on Pro and Claude Opus 5.5 on Ultra. Credits reset monthly and do not roll over. When they run out you can wait for the reset, upgrade or buy a one-time credit pack. Web search is included and uses about one credit per search from the same monthly balance. The pages a search brings back still add to the message's length. A short per-minute safety limit stops automated abuse. Normal use does not reach it.

Estimated messages per month

Assumes a short message of about 1,500 tokens in and 700 out. Long chats, files, web results and heavy reasoning use more. Free (Claude Haiku 5.5: About 1,000; Claude Sonnet 5.5: Not on Free; GPT-6.1 Sol: Not on Free; Claude Opus 5.5: Not on Free). Go ($8) (Claude Haiku 5.5: About 12,000; Claude Sonnet 5.5: Not on Go; GPT-6.1 Sol: Not on Go; Claude Opus 5.5: Not on Go). Pro ($15) (Claude Haiku 5.5: About 20,000; Claude Sonnet 5.5: About 1,000; GPT-6.1 Sol: About 1,000; Claude Opus 5.5: Not on Pro). Ultra ($40) (Claude Haiku 5.5: About 50,000; Claude Sonnet 5.5: About 2,500; GPT-6.1 Sol: About 2,500; Claude Opus 5.5: About 1,250). Max ($200) (Claude Haiku 5.5: About 250,000; Claude Sonnet 5.5: About 12,500; GPT-6.1 Sol: About 12,500; Claude Opus 5.5: About 6,250). The spread is wide because model prices are. A Claude Opus 5.5 message costs about 40 times as much as a Claude Haiku 5.5 message. Model prices come from Quad's model catalog, and the LLM pricing comparison lists every one. Your remaining credits show in Settings, and Quad warns you before they run out.

Images and video per month

Estimates if you spent the whole month's budget on one model, at the setting named in each column. Free (GPT Image 1 Mini images: About 100; Nano Banana Pro 1 images: Not on this plan; Veo 3.1 Lite clips (4 s, 720p): Not on this plan). Go ($8) (GPT Image 1 Mini images: About 1,200; Nano Banana Pro 1 images: Not on this plan; Veo 3.1 Lite clips (4 s, 720p): Not on this plan). Pro ($15) (GPT Image 1 Mini images: About 2,000; Nano Banana Pro 1 images: About 74; Veo 3.1 Lite clips (4 s, 720p): Not on this plan). Ultra ($40) (GPT Image 1 Mini images: About 5,000; Nano Banana Pro 1 images: About 186; Veo 3.1 Lite clips (4 s, 720p): About 208). Max ($200) (GPT Image 1 Mini images: About 25,000; Nano Banana Pro 1 images: About 930; Veo 3.1 Lite clips (4 s, 720p): About 1,041). Images and video come out of the same monthly budget as chat. Video starts on Ultra. The AI video generator page lists the video models and the plan each one needs.

Without an account

You can send 3 messages a day at quad.chat without signing up. That is enough to see how the app answers. A free account adds saved chats, the monthly Free budget, web search and the iPhone and Android apps, with no credit card.

How this compares with ChatGPT and Claude limits

OpenAI and Anthropic set their own limits and change them often, so we do not quote numbers we cannot keep current. In general, ChatGPT plans cap how many messages you can send to certain models within a time window, and Claude plans meter usage over rolling session windows with weekly limits on top. Check each vendor's help center for the figures that apply today. The practical difference is what happens when you reach the limit. In a single-vendor app you wait. In Quad you can move the same conversation to a cheaper model and keep going on what is left of the month's budget. If you mostly use one premium model all day, every day, that vendor's own top plan may still be better value, and the Claude alternative page covers that trade-off.

Frequently asked questions

How many messages do I get on the Free plan?

About 1,000 short Claude Haiku 5.5 messages a month, or a similar number on other low-cost free models such as GPT-6 Luna. Longer messages and files use more. Web search is included and draws from the same monthly credit balance.

What is the Claude Opus 5.5 message limit in Quad?

There is no fixed Opus cap. Opus 5.5 is available from Ultra, and a full Ultra month buys about 1,250 short Opus messages if you use nothing else. Mixing in cheaper models stretches the budget further.

How many Claude Sonnet 5.5 messages does Pro include?

About 1,000 short messages a month if you use only Sonnet 5.5. GPT-6.1 Sol costs the same per token in Quad's catalog, so the estimate is the same for it.

Is there a per-model or per-day message cap?

No. Quad uses one monthly budget across every model your plan includes. A short per-minute safety limit blocks automated abuse, and guests without an account get 3 messages a day.

What happens when I run out of messages?

You can switch to a cheaper model to stretch what is left, buy a one-time credit pack, upgrade, or wait for the monthly reset. Unused credits do not roll over.