Download
04 · AI Models

Three tiers. You pick.

Everyday drafting uses the Apple foundation model at the core of Apple Intelligence — on your iPhone, built in, instant, even offline. With iOS 27, Apple's Private Cloud Compute steps in when a turn needs more room than the phone has: the same Apple Intelligence stack, on Apple silicon servers, with a much larger context window and deeper reasoning. And when the work is genuinely long and multi-step, Claude takes over on prepaid AI Credit or your own Anthropic key. You switch in the AI Assistant chat itself, or per feature in Settings › AI Features.

On your iPhone when…

It's an everyday reply, you want it instantly, you're offline, or nothing should leave your phone. In the AI Assistant chat you can pin this tier from the model picker; the other features move between the two Apple tiers on their own.

Private Cloud Compute when…

The thread is longer than the phone can hold at once, or you want your Knowledge Base documents searched — still Apple Intelligence, still no key, just more room to think.

Hand it to Claude when…

The task has steps, you need research with sources, a document produced, or a nuanced tone for a message that matters.

CapabilityApple IntelligenceClaude
Where it runsTwo tiers, both Apple. On-device, powered by the foundation model at the core of Apple Intelligence — no network required, works offline. Or Apple's Private Cloud Compute on iOS 27, the same stack on Apple silicon servers, for the turns that need more room. You pick which in the chatAnthropic cloud — with prepaid AI Credit, or direct to Anthropic on your own API key
What you payBuilt in — no API key needed. The Free plan includes 5 Apple Intelligence actions a day (shared across Smart Reply, Chatbot, Unread Summary, Skills, AI Writer, Split Bill, Poll from Photo, poll options and the AI Assistant); every paid plan makes them unlimited. Writing Tools is never countedTwo ways in: AI Credit — prepaid top-ups, no subscription and no API key needed — or any paid plan (Solo, Duo or Pro) plus your own key, paying Anthropic directly per use with no markup
ModelsApple Foundation Models — the ~3B on-device model, or Apple's larger server-side model on Private Cloud ComputeOpus 5.5 (most capable for ambitious work), Sonnet 5.5 (most efficient for everyday tasks), Haiku 5.5 (fastest for quick answers)
Context window4,096 tokens on device (input + output combined, ~3,000 words); about 32,000 on Private Cloud Compute. Longer threads are condensed automatically either wayUp to 1,000,000 tokens on every model — handles entire conversation histories. The same on AI Credit and on your own key
Input typesText only — images, PDFs, and documents are handled by ClaudeMultimodal — understand and analyse a wide range of visual formats, including photos, charts, graphs, and technical diagrams
Extended thinkingNot on the on-device model, which is tuned for direct, single-pass responses. Private Cloud Compute brings the deeper reasoning the larger model allowsEnhanced reasoning capabilities for complex tasks — Claude's step-by-step thought process is visible before it delivers its final answer, with adaptive thinking and adjustable effort (Low, Medium, High, Extra, Max)
Reasoning and logicOn device, best for simple focused tasks like everyday drafting rather than heavy logic or math. Private Cloud Compute takes the heavier ones — more context held at once, and reasoning to go with itStrong complex reasoning, multi-step problem solving, code generation, and mathematical tasks
Tool useNearly all of them, running through Apple's own frameworks — web search, URL fetching, contacts, messages and chat discovery, calendar, reminders, location, files, and image generation. Knowledge Base document search needs Private Cloud Compute; on the on-device model only saved links can be opened. The Claude-native ones — Task Tracking, Memory, Sub-agents, Code Execution, and Agent Skills — sit outside the Apple setThe full set. Claude decides when to call a tool from your request and the tool's description, then AI Reply Assistant executes it on your device. It adds the agent-only tools on top of everything the on-device set already has. Files and Agent Skills file creation (PPTX, XLSX, PDF, DOCX) need your own Anthropic key
Agentic behaviourSingle-turn — one prompt, one responseClaude dynamically directs its own processes and tool usage, maintaining control over how it accomplishes tasks — planning, calling tools, observing results, and refining across multiple iterations
Document creationNot availablePre-built Agent Skills let Claude run code in a sandboxed container to create PowerPoint presentations, Excel spreadsheets with charts, Word documents, and PDF reports — downloaded and delivered as shareable attachments. Requires your own Anthropic key
Message draftingGenerates up to 3 drafts in different tones — you pick the one to use before anything is sentHuman-in-the-loop — Claude drafts messages as cards, you review, choose recipients, and confirm before anything is sent
Context managementWhen a thread outgrows the context window, the transcript is automatically condensed and the conversation continues — no error, nothing lostCompaction extends effective context length by automatically summarising older context, keeping the active context focused and performant. Prompt caching reduces processing time and costs for repetitive tasks. On-demand tool loading and real-time token tracking
Languages9 languages (English, French, German, Italian, Portuguese, Spanish, Japanese, Korean, Chinese)Broad multilingual support across dozens of languages
Factual knowledgeA compact on-device model — tuned for everyday drafting rather than broad factual recall or current eventsBroader world knowledge with more current training data — stronger on factual and domain-specific questions
PrivacyRuns on your iPhone, or on Apple's Private Cloud Compute when a turn needs more room. Nothing goes to Anthropic or to usUsed only for the features you run on it. What a feature works on is sent to Anthropic — directly on your own key, or through Flowbie's server, which does not store the content, with AI Credit.
Best forEveryday replies, quick suggestions, full privacy, offline useEven more expertise — longer drafts, nuanced tone, complex reasoning, multi-step tasks, document creation, and research

Which Claude, specifically

All three, on either route into Claude — models are never gated by plan or by how you pay, and the context column holds either way, so AI Credit runs every model at its full native window exactly as your own key does. Switch mid-conversation and the next response uses your selection. All three support Extended Thinking. We move to Anthropic’s newest release in each tier as it ships, so these are the current models rather than a fixed list.

ModelStrengthContextBest for
Haiku 5.5Fastest, most cost-efficient1MQuick drafting
Sonnet 5.5Best balance of speed + intelligence1MEveryday and complex tasks
Opus 5.5Most capable for deep reasoning1MComplex, long-running work

Which AI runs each feature — Settings › AI Features

Settings › AI Features lists the AI features in the chat — Smart Reply, Agent, Chatbot, AI Writer, Writing Tools, poll and checklist options, Split Bill, Poll from Photo, and Unread Summary. Each one has a Provider row that picks Apple Intelligence or Claude. While Claude runs a feature, a Model row shows its Claude model: Smart Reply, Agent, Chatbot, AI Writer, Split Bill and Poll from Photo let you pick Opus 5.5, Sonnet 5.5 or Haiku 5.5, and the other three run on Sonnet 5.5. You can pick Claude before it is set up: the app opens the Claude screen, and the feature switches to Claude once you add AI Credit or your own Anthropic key. You can switch any feature back to Apple Intelligence at any time.

On an iPhone or iPad without Apple Intelligence, the Provider row reads Claude, because every feature runs on Claude there. Until Claude is set up, the Provider and Model rows are greyed, and tapping one offers to set Claude up. On the Apple side there is nothing to configure: a turn runs on Private Cloud Compute when it can serve and on device when it cannot, and in the AI Assistant chat you can pin the tier yourself from the model picker.

Writing for each engine

The table above is what each one can do. This is what that means for the instructions you write in a Skill: the same prompt does not get the best out of both.

When you writeApple IntelligenceClaude
Prompt lengthKeep it short — 2–4 sentences. The on-device model has a 4,096-token context window (input + output combined), so every word counts.Can handle longer, more detailed prompts — up to 1M tokens on Sonnet 5.5, Opus 5.5 and Haiku 5.5, whichever way you pay. Add examples, edge cases, and nuanced instructions.
ExamplesUse 1–2 simple examples. Too many or too complex will consume the limited context and cause parroting.Use 3–5 diverse examples for best results. Claude generalises well from richer input.
Conditional logicKeep conditionals simple. The on-device model handles if-this-then-that branching poorly — reach for the tone sliders instead of writing rules.Can handle moderate conditional complexity — "If the message is a complaint, be empathetic; if it's a question, be direct."
ReasoningDon't ask it to think step by step. At this size the reasoning tends to leak into the reply itself, and the logic is unreliable anyway.Enable Extended Thinking for complex tasks. Claude handles chain-of-thought reasoning natively with visible, adjustable thinking.
Persona depthA single sentence for role + persona is usually enough. Longer persona prompts eat into the 4K token budget.You can build richer personas with backstory, communication preferences, and relationship dynamics.
Images and documentsNot supported — the on-device model is text-only. No images, PDFs, or file attachments.Include screenshots, photos, charts, and PDFs. Claude analyses visual content alongside text for richer context.

Which entry uses which is set per Skill, so a quick everyday Prompt and a research-heavy Skill can run on different engines — templates and the mistakes to avoid are on the Skills page.

What it costs

Apple Intelligence runs on every plan with no per-message fees — the Free plan includes 5 Apple AI actions a day (Smart Reply, Chatbot, Unread Summary, Skills, AI Writer, Split Bill, Poll from Photo, poll options and the AI Assistant), and every paid plan makes them unlimited. Writing Tools and anything that runs on Claude never count toward them. That covers both Apple tiers: Private Cloud Compute costs nothing extra here, and is metered by Apple against your own device's allowance rather than by us — if you reach it, the app says so and keeps working on device. Claude comes two ways. AI Credit works on any plan: buy a prepaid top-up (from US$4.49), sign in with Apple, and spend it as you go — no subscription, no API key, and it never expires. Or subscribe — Solo US$19.99/year or US$2.99 monthly, Duo US$32.99/year or US$4.99 monthly, Pro US$49.99/year or US$7.49 monthly, each with a 7-day free trial — and bring your own Anthropic API key: you pay Anthropic's standard usage rates directly with no markup, can cap spending in your Anthropic console, and add Managed Agents, Agent Skills, and Files. Bringing a key works on all three plans; what they differ on is how many messaging accounts you connect at once. Context is the same either way — each model runs at its full native window, so Sonnet 5.5 and Opus 5.5 give you 1M tokens whether you paid with credit or a key. Subscribing also puts AI Credit in your balance once your first bill goes through — US$1.00 on Solo, US$1.50 on Duo, US$2.00 on Pro, once per Apple Account — so you can try Claude without buying a top-up. Either way, a typical drafted reply uses only a few thousand tokens, and you choose the model per task — Haiku for fast, inexpensive drafting up to Opus for the hardest work.

Compare Free, Solo, Duo and Pro →

Two ways to run Claude

AI Credit gets anyone started in a minute — no subscription, no Anthropic account. Bringing your own key is the full-surface tier: you deal with Anthropic directly, and Managed Agents, Agent Skills, and Files come with it. Context is not part of the difference — every model runs at its full native window on either route.

AI Credit versus your own API key comparison
CompareAI Credit
Prepaid top-ups · any plan
Your own API key
Any paid plan · direct to Anthropic
Who it’s forAnyone — no subscription, no Anthropic account, no API keySubscribers on any paid plan who want the full Claude surface and a direct Anthropic relationship
SetupSign in with Apple, buy a top-up, start chattingCreate an Anthropic account, mint an API key, paste it in once
BillingPrepaid credit in dollars, spent as you use Claude. Never expires; pack prices include the cost of running the model. Any balance above zero starts a request — there is no minimum to beginYou pay Anthropic directly at their standard usage rates — no markup and no per-message fee from us
ModelsHaiku, Sonnet, and Opus — never gated by planHaiku, Sonnet, and Opus
Context windowUp to 1,000,000 tokens (Opus and Sonnet)Up to 1,000,000 tokens (Opus and Sonnet)
AI features in a chatSmart Reply, Agent, Chatbot, AI Writer, Writing Tools, poll options, Split Bill, Poll from Photo and Unread Summary — each one you choose Claude for in Settings › AI Features, and all of them on a device without Apple IntelligenceSmart Reply, Agent, Chatbot, AI Writer, Writing Tools, poll options, Split Bill, Poll from Photo and Unread Summary — each one you choose Claude for in Settings › AI Features, and all of them on a device without Apple Intelligence
Agent chat & AI ToolsAgent chat included, with every tool except two — the Files API and Agent Skills both need your own keyAgent chat included, with every tool
Web searchIncluded — billed from credit per searchIncluded — billed by Anthropic per search
Managed AgentsNot availableIncluded — server-side sessions that keep working across devices
Agent Skills — PPTX, XLSX, PDF, DOCXNot available — generating these files runs in Anthropic's code container, which is billed to an Anthropic organisation rather than to prepaid creditIncluded — PowerPoint, Excel, PDF, and Word generated in Anthropic’s code container
Privacy pathThrough Flowbie’s server — content is not stored; logs carry only billing metadataStraight to Anthropic under your own account — Flowbie is never in the path
If you have bothYour balance sits untouched — it is not spent while a usable key is connectedYour key serves every request, so you are never charged twice for the same reply

What turns Claude on

Claude needs one of two things: prepaid AI Credit, or your own Anthropic key on a paid plan. A subscription by itself does not include Claude usage — here is every combination.

Which combination of plan, AI Credit, and API key enables Claude
Your setupWhat you get
Free plan · no AI CreditClaude is off. On a device with Apple Intelligence, it still runs Smart Reply, Skills, Chatbot, Unread Summary, AI Writer, Split Bill, Poll from Photo and poll options, up to 5 actions a day, and Writing Tools without a limit. On a device without Apple Intelligence, the AI features run once you add AI Credit; until then, tapping one offers to set Claude up.
Free plan · AI CreditClaude is on. The agent chat, any AI feature you choose Claude for in Settings › AI Features — every one of them on a device without Apple Intelligence — and every tool except the two that need your own key. No subscription needed.
Paid plan · no key, no AI CreditClaude is off. A paid plan on its own does not include Claude usage — add AI Credit or connect your own Anthropic key. The Apple Intelligence actions the Free plan caps at 5 a day are uncapped on every paid plan. On a device without Apple Intelligence, the AI features need AI Credit or your own key.
Paid plan · AI CreditClaude is on, spent from your prepaid balance. Same surface as the Free plan with credit.
Paid plan · your own keyClaude is on, billed by Anthropic to your own account. Adds Managed Agents, Agent Skills, and Files. Available on Solo, Duo and Pro alike — bringing a key is not what the tiers differ on.
Paid plan · your own key · AI CreditYour key serves every request and your credit is left untouched — you are already paying Anthropic, so we never spend the balance on top.

On AI Credit you pick any of the three models at its full context window — the model list is never gated by how you pay. Requests are metered against your balance as they run, and when it is spent Claude asks you to top up.

Your personal AI assistant for messaging.

Download AI Reply Assistant on iOS and try it free.