The latest AI models. One app.
When a new AI model comes out, you get it — automatically.
Generous everyday chat with Use AI Instant
Limited web search
Web, iOS & Android
Fair use applies. Learn more
Unlimited messaging with Use AI Instant
Premium AI models
Projects & knowledge bases
Switch models mid-chat
Image generation with the latest AI models
Deep Research
Create documents, slides & sheets
Connect 100+ apps
Usage limits apply. Learn more
Claude Fable 5.1
5× the usage allowance
Maximum reasoning
Maximum Deep Research
Faster image generation
Maximum memory & context
Early access to new features
Usage limits apply. Learn more
Compare features across plans
| Essentials | Pro |
| Everyday text chats | Unlimited[1] |
| Premium model usage | Included[1] |
| Usage Boost | Coming soon |
| Response times | Fast[2] |
| Chat history | Unlimited |
| Access on Web, iOS & Android | |
| Memory & context | Standard[2] |
| Early access to new features |
| Chat | Pro |
| Use AI Instant | |
| Switch models mid-chat | |
| Premium AI models | |
| Claude Opus 5.5 | |
| Claude Sonnet 5.5 | |
| Claude Fable 5.1 | |
| GPT-5.6 Sol | |
| GPT-5.5 | |
| GPT-5.4 | |
| GPT-6 Astra | |
| Gemini 3.8 Flash | |
| DeepSeek V4 Pro | |
| Kimi K3 | |
| Kimi K2.6 | |
| Grok 4.7 | |
| GLM 5.2 | |
| New models as they come out | |
| Reasoning effort | High[3] |
| Create | Pro |
| Documents, PDFs, slides & sheets | |
| Image generation with leading AI models | |
| Nano Banana 2 | |
| Nano Banana 2 Lite | |
| GPT Image 2 | |
| FLUX.2 Pro | |
| MAI-Image-2.5 | |
| Krea 2 Medium Turbo | |
| Krea 2 Medium | |
| Research & tools | Pro |
| Web search | |
| Deep Research | Up to 200 sources |
| 100+ app connectors | |
| AI agents | |
| Projects | |
| Knowledge bases | |
| File library |
[1] Each plan includes a monthly usage allowance for premium AI models. The limit is applied weekly on paid plans — and daily during a trial — proportional to your monthly allowance. Messaging with Use AI Instant is unlimited on paid plans, subject to fair use. Learn more. See also the Fair Usage Policy.
[2] Response times and memory & context vary by model — per-plan context windows apply where the model provider supports them, and each model has its own speed and maximum context window.
[3] Reasoning effort is how long a model thinks before it answers — more effort means more careful answers to complex questions: better math, code, and analysis. Free runs at low effort, Pro at high, and Max at maximum. Models without adjustable reasoning run at their default setting.
[4] For faster image generation, we ask the model provider to prioritize your request — prioritization itself is not guaranteed.
How many messages does a plan buy?
Here's what each plan buys, model by model — and where unlimited kicks in.
| Pro | Max | |
|---|---|---|
| Use AI InstantNever counts toward your usage | Unlimited | Unlimited |
| Claude Opus 5.5 | ≈ 952 / mo≈ 238 / wk | ≈ 4,761 / mo≈ 1,190 / wk |
| Claude Sonnet 5.5 | ≈ 1,904 / mo≈ 476 / wk | ≈ 9,523 / mo≈ 2,380 / wk |
| Claude Fable 5.1 | Max only | ≈ 1,904 / mo≈ 476 / wk |
| GPT-5.6 Sol | ≈ 952 / mo≈ 238 / wk | ≈ 4,761 / mo≈ 1,190 / wk |
| GPT-5.5 | ≈ 666 / mo≈ 166 / wk | ≈ 3,333 / mo≈ 833 / wk |
| GPT-5.4 | ≈ 1,333 / mo≈ 333 / wk | ≈ 6,666 / mo≈ 1,666 / wk |
| GPT-6 Astra | ≈ 380 / mo≈ 95 / wk | ≈ 1,904 / mo≈ 476 / wk |
| Gemini 3.8 Flash | ≈ 5,079 / mo≈ 1,269 / wk | ≈ 25,396 / mo≈ 6,349 / wk |
| DeepSeek V4 Pro | ≈ 4,040 / mo≈ 1,010 / wk | ≈ 20,202 / mo≈ 5,050 / wk |
| Kimi K3 | ≈ 1,269 / mo≈ 317 / wk | ≈ 6,349 / mo≈ 1,587 / wk |
| Kimi K2.6 | ≈ 4,519 / mo≈ 1,129 / wk | ≈ 22,598 / mo≈ 5,649 / wk |
| Grok 4.7 | ≈ 2,666 / mo≈ 666 / wk | ≈ 13,333 / mo≈ 3,333 / wk |
| GLM 5.2 | ≈ 3,703 / mo≈ 925 / wk | ≈ 18,518 / mo≈ 4,629 / wk |
Out of premium usage? Switch to Use AI Instant and keep chatting — messaging with it is unlimited on Pro and Max, subject to fair use.
Estimates only — actual counts vary with conversation length, attachments, reasoning effort and web search. Fair usage limits apply. Learn more.