How to Change Model, Effort & Extended Thinking in the Claude App

When you open a conversation in claude.ai, you'll notice a small button next to the send field showing the current model name. That single button is your control panel for which Claude model you're using, how deeply it thinks (effort), and whether it shows its reasoning (extended thinking). No developer tools or API — a way of connecting to Claude through programming code — required. This guide covers everything through the claude.ai web interface, step by step.

Update (September 27, 2026): Claude Opus 5.5 was released on September 22, 2026 and is now the current Opus model (API claude-opus-5-5, $4 per million input tokens and $20 per million output tokens, down from Opus 5's $5/$25). Anthropic says it performs at the level of Fable 5.1 on most work and costs 40% less to run than Opus 5 on typical workloads. Opus 5 remains available but is marked legacy in the official model docs. Opus 5 prices and comparisons below reflect the pre-launch state and will be updated. Details: Claude Opus 5.5 complete guide.

Before You Start

This guide is written for the claude.ai website — not a mobile app or API. Please check the following before diving in.

• Account: You need to be signed in to claude.ai. Basic model switching works on the Free plan.
• Plan differences: Effort levels, Extra High, and Max options are more broadly available on Pro, Max, Team, and Enterprise plans. Free plan users may not see all options.
• Enterprise note: If your organization's administrator has disabled certain models or effort levels, those options will simply not appear in the menu. Contact your admin if something you expect is missing.

① Switching Models — Opus, Sonnet, and Haiku

Claude comes in three model families, each suited to different tasks. Based on the official model overview (as of August 2026), the current models are:

• Claude Opus 5 — The most capable option for complex reasoning, in-depth research, and demanding coding tasks.
• Claude Sonnet 5 — A well-balanced model for everyday writing, summarization, and general questions.
• Claude Haiku 4.5 — Light and fast, ideal when you need a quick reply for simple translations or short Q&A.

How to switch models

1. Start a new conversation or open an existing one in claude.ai.
2. Find the model name displayed next to the send button. Click it.
3. Choose a different model from the dropdown. Click "More models" to see additional options.

✅ Success signal: The model name next to the send button updates to the one you selected. The change takes effect from your very next message — you can switch mid-conversation at any time.

② Effort Levels — Choosing How Hard Claude Works

Effort controls how much processing power Claude applies to each response. Think of it like choosing between a quick note and a fully researched report. Higher effort means more thorough answers, but they take longer and consume your usage allowance faster. According to the official documentation, the effort selector is currently available for models including Opus 5 and Sonnet 5.

How to change the effort level

1. Click the model name next to the send button.
2. Click "Effort" in the menu.
3. Select your preferred level: Low, Medium, High, Extra High, or Max.

✅ Success signal: The selected effort level appears in the menu or next to the model name. The change takes effect from your next response.

⚠️ Watch out: Extra High and Max produce better answers, but they drain your usage allowance much faster. If you're watching your limits, use Low or Medium for routine tasks and save higher levels for work that genuinely needs them. According to the official documentation, Extra High is available on recent Opus models such as Opus 5 and Opus 5.

③ Extended Thinking — Seeing Claude's Reasoning

Extended thinking — the Thinking toggle — lets Claude show you its step-by-step reasoning before it gives a final answer. Imagine a tutor who not only tells you the answer but writes out every step on the whiteboard. Thinking and effort are separate settings: you can run Low effort with thinking on, or Max effort with thinking off — any combination works.

For models with the effort selector (such as Opus 5):

1. Click the model name next to the send button.
2. Mouse over "Effort".
3. Switch the "Thinking" toggle on or off.

For other models:

1. Click the model name next to the send button.
2. Switch the "Extended" toggle on or off.

✅ Success signal: With thinking enabled, you'll see a "Thinking" indicator with a running timer while Claude processes your message. Once Claude responds, an expandable "Thinking" section appears above the answer. Click it to read through Claude's reasoning summary.

⚠️ Watch out: Occasionally you may see a message saying the rest of Claude's thought process isn't available — this means the safety system stepped in. According to the official documentation, this happens when the reasoning touches content identified as potentially harmful. Try rephrasing your question and sending it again.

Which Settings to Use and When

Based on the official documentation's guidance, here's a practical starting point:

• Casual questions, short writing, quick lookups: Default settings (High effort, thinking off) are plenty. You'll also preserve more of your usage allowance.
• Math problems, proofs, competition-level coding: Raise the effort and turn on extended thinking.
• Comprehensive project planning, detailed document analysis: Try High or above with thinking enabled.
• Complex coding and agentic tasks — tasks where Claude takes multiple automated steps: Start with Extra High on Opus 5. If the result still isn't quite right, step up to Max.
• Work where accuracy is non-negotiable: Max effort, and expect a longer wait for the response.

Regardless of the settings you pick, specific prompts outperform vague ones every time. "Find any conflicting terms between Clause 3 and Clause 7 of this contract" will get better results than "check this contract" — even at the same effort level.

Practical Tips to Save Your Usage Allowance

• Use Low or Medium for everyday chat: Quick translations, short summaries, and simple lookups don't need high effort. Your allowance will last much longer.
• Reserve Max for what actually needs it: Save the most thorough setting for important analyses or hard problems — not routine follow-up questions.
• You can change settings mid-conversation: Start at High, and if the answer isn't detailed enough, switch to Max right there and ask again. No need to start a new conversation.
• Turn thinking off when you don't need to audit the reasoning: Extended thinking slows responses slightly. Enable it when you want to verify how Claude arrived at a conclusion, or when working through genuinely complex multi-step problems.

Common Trouble Spots and Fixes

• The Effort menu doesn't appear: The currently selected model may not support effort levels. Switch to Opus 5 or another supported model. A Free plan may also limit access to this option.
• Extra High or Max is missing: Your plan or an Enterprise administrator setting may have disabled those options. Enterprise users should contact their admin.
• No thinking toggle visible: The selected model may not support extended thinking. Try switching models, or check whether the Thinking toggle is nested inside the Effort submenu.
• Responses are taking too long: Max effort or extended thinking increases processing time. If speed matters more right now, lower the effort level and turn thinking off.
• Thinking stopped partway through: The safety system intervened. Rephrase your request and try again.

Next Steps

Once you're comfortable with model, effort, and thinking settings, the next big lever is writing better prompts — the instructions you send to Claude. Clearer, more specific prompts amplify whatever settings you've chosen. You might also explore claude.ai's Projects feature, which lets you save context and preferences so you don't have to reconfigure things every time you start a new conversation.

Related articles

• Claude Privacy and Data Settings: Are My Chats Used for Training?
• Claude Settings Explained — Every Option and What to Turn On (2026)
• Claude Appearance Settings Guide — How to Switch Dark/Light Mode and Change Chat Font
• Claude Projects, Artifacts & Memory: What They Are and How to Use Them
• How to Write Prompts for Claude: 6 Core Principles from the Official Docs
• How to Adapt When a Claude Model Is Deprecated — Official Guide