NEW · Morning journal prompts → start your day with intention
Random Prompts

Anthropic — New Default Model June 30, 2026

Claude Sonnet 5 Prompt Generator

20 free Claude Sonnet 5 prompts across reasoning modes, coding, writing, analysis, and agentic tasks — with selectable effort levels. Copy directly into Claude.ai. No signup.

1M Token Context Window Selectable Reasoning: Low → xhigh Near-Opus 4.8 Quality New Claude.ai Default $2/$10 per MTok

What Is the Claude Sonnet 5 Prompt Generator?

The Claude Sonnet 5 prompt generator on this page gives you 20 copy-paste prompts designed to unlock the key capabilities of Anthropic's newest default AI model. Claude Sonnet 5 (model ID: claude-sonnet-5) launched June 30, 2026 and immediately replaced Sonnet 4.6 as the default model across Claude.ai and GitHub Copilot. It delivers near-Opus 4.8 performance at Sonnet pricing — $2/$10 per million tokens — with a 1M token context window and a first-of-its-kind selectable reasoning effort system.

The prompts on this page are organised by reasoning effort level (low, medium, high, xhigh) and use case. Each one shows you how to specify the right effort level for the task — the biggest mistake new Sonnet 5 users make is using high-effort reasoning on simple tasks (slow and expensive) or low-effort on complex ones (shallow output). Paste any prompt directly into Claude.ai and fill in the bracketed placeholders with your specific context.

Claude Sonnet 5 Reasoning Effort Levels — When to Use Each

Low
Fast & Cheap

Simple rewrites, formatting, lookups, one-sentence answers, tone adjustments. No heavy reasoning needed.

Medium
Balanced

Writing, standard code review, data summaries, email drafts, API docs. The right default for most tasks.

High
Deep Reasoning

Research synthesis, complex debugging, strategy, PRDs, multi-step agentic tasks, long-form articles.

xhigh
Extended Thinking

Hard logic puzzles, mathematical proofs, root cause analysis on complex bugs, architectural decisions.

20 Free Claude Sonnet 5 Prompts — Copy & Paste

Each prompt shows the recommended reasoning effort level. Works with Claude.ai free, Pro, and the Anthropic API.

R
Reasoning: Low effort: low

Reasoning Mode — Quick Analysis

Use low reasoning effort. Summarise the key argument in this paragraph in one sentence and identify any logical gaps: [Paste your paragraph]

C
Reasoning: Medium effort: medium

Code Review — Spot the Bug

Use medium reasoning effort. Review this function for bugs, edge cases, and performance issues. Return: (1) a bullet list of all problems found, (2) corrected code, (3) one-line explanation of each fix. ``` [Paste your code here] ```

R
Reasoning: High effort: high

Research Paper Synthesis

Use high reasoning effort. I am attaching three research papers on [topic]. Synthesise them into a structured report: 1. CONSENSUS — what all three papers agree on 2. CONTRADICTIONS — where they disagree, and why the disagreement exists 3. GAPS — what none of the papers address 4. TAKEAWAY — one-paragraph practical conclusion for a non-specialist audience [Paste paper text or excerpts]

H
Reasoning: xhigh effort: xhigh

Hard Logic Puzzle

Use extended thinking (xhigh reasoning effort). Solve this step by step, showing every deduction: Five competitors sit in a row. Each has a different nationality, drink, sport, pet, and car brand. Use these clues to determine who owns the fish: [Paste your logic puzzle clues here] Show each deduction and elimination explicitly before stating the answer.

F
Coding effort: high

Full Stack Feature — Spec to Code

Build a complete, production-ready implementation of the following feature: Feature: [describe your feature] Stack: [your tech stack, e.g. Next.js 15 + TypeScript + PostgreSQL + Prisma] Deliver: 1. Database schema changes (if any) with migration SQL 2. API route(s) with input validation and error handling 3. React component with loading, error, and empty states 4. Unit test stubs for each new function 5. A brief comment in the code explaining any non-obvious decisions Follow the existing patterns in the codebase: [describe your patterns or paste a sample file]

A
Agentic effort: high

Agentic Task — Multi-Step Plan

Plan and execute the following multi-step task. At each step, tell me what you are doing, what information you need, and what decision you are making before proceeding: Task: [describe your goal, e.g. 'Analyse our Q2 sales data, identify the top 3 underperforming product lines, and draft a concise brief for the product team with specific action recommendations.'] Available context: - [Paste your data, documents, or context] After completing all steps, write a final one-paragraph summary of what you found and what I should do next.

W
Writing effort: medium

Writing Coach — Paragraph Rewrite

Rewrite the following paragraph three times, each with a different goal: 1. VERSION A — tighter and clearer: cut every word that doesn't earn its place. Target 30% shorter. 2. VERSION B — more persuasive: restructure to lead with the strongest claim and end with a call to action. 3. VERSION C — more authoritative: raise the register, remove hedging language, and add one concrete data point placeholder where appropriate. For each version, add one sentence explaining the key change you made. Original paragraph: [Paste your paragraph]

A
Coding effort: medium

API Documentation Draft

Write complete API documentation for the following endpoint in OpenAPI 3.1 YAML format, then also write a human-readable markdown reference section. Endpoint details: - Method: [GET/POST/etc] - Path: [/your/endpoint] - Auth: [bearer token / API key / none] - Request body: [describe fields] - Success response: [describe fields] - Known error codes: [list them] The documentation should be clear enough for a developer who has never used this API before to make a successful request on the first try.

D
Analysis effort: high

Decision Framework

Help me make the following decision using a structured decision framework: Decision: [describe your decision, e.g. 'Should we build this feature in-house or buy a third-party tool?'] Context: - Constraints: [time, budget, team size, etc.] - Stakeholders: [who is affected] - Options under consideration: [list them] Deliver: 1. A weighted decision matrix with the 5 most important criteria 2. Your recommendation with a clear rationale 3. The strongest argument AGAINST your recommendation (steelman the losing option) 4. What new information would change your recommendation

E
Writing effort: low

Email Tone Fix

Rewrite this email to be professional but warm — direct without being cold, confident without being arrogant. Keep the core message identical. Do not add filler phrases like 'I hope this email finds you well.' Flag any sentence that could be misread as passive-aggressive and suggest an alternative. [Paste your email draft]

S
Coding effort: medium

SQL Query Optimiser

Analyse this SQL query for performance issues and rewrite it to be faster. Explain each optimisation. Database: [PostgreSQL / MySQL / SQLite / other] Table sizes: [approximate row counts, e.g. 'orders: 12M rows, users: 800K rows'] Current query: ```sql [Paste your query] ``` Provide: 1. The rewritten query 2. Bullet list of every change made and why it helps 3. Which indexes to add (if any) with the exact CREATE INDEX statement

C
Analysis effort: high

Content Strategy Audit

Audit my website's content strategy based on the following information and give me a prioritised action plan. Site overview: [describe your site, audience, and goal] Top 5 pages by traffic: [list them with approximate monthly visits] Top 5 keywords ranking on page 1: [list them] Top 5 keywords ranking on page 2-3: [list them — these are your quick-win opportunities] Content published last 90 days: [list titles] Deliver: 1. A diagnosis: what is the content strategy currently optimised for? 2. The 3 biggest gaps between what you publish and what your audience searches for 3. A prioritised content calendar for the next 8 weeks (specific titles, not categories) 4. The one page I should update immediately and why

S
AI Engineering effort: high

System Prompt Design

Write a production-quality system prompt for a Claude Sonnet 5 deployment with the following requirements: Use case: [describe the bot, e.g. 'customer support agent for a B2B SaaS tool'] Tone: [formal / casual / technical / empathetic] Out-of-scope handling: [what should it refuse or redirect?] Knowledge boundaries: [what does it know and not know?] Output format rules: [how should it structure responses?] The system prompt must: - Clearly define persona and role in the first 3 sentences - Include explicit fallback instructions when the model doesn't know the answer - Add a 'reasoning effort' instruction appropriate to the task complexity - Be under 800 tokens Also write 3 test prompts I can use to verify the system prompt is working correctly.

C
Analysis effort: medium

Competitive Landscape Brief

Write a competitive landscape brief for the following market: Our product: [one sentence description] Our target customer: [describe] Competitors to analyse: [list 4-6 companies] For each competitor provide: - Their primary value proposition (one sentence) - Estimated customer profile (who buys it) - One thing they do better than us - One thing we do better than them - Their apparent pricing model End with: the white space in this market — what positioning is genuinely unclaimed that we could own?

T
Coding effort: medium

Test Suite Generator

Generate a complete test suite for the following function using [Jest / Vitest / pytest / your framework]. Cover: happy path, edge cases, invalid inputs, boundary conditions, and any async behaviour. Function: ``` [Paste your function here] ``` For each test: - Name it descriptively (what scenario + what result) - Use the AAA pattern (Arrange / Act / Assert) - Add a one-line comment explaining why this case matters

L
Writing effort: high

Long-Form Article Draft

Write a 1,500-word article on the following topic for [your publication / audience]: Topic: [your topic] Primary keyword (must appear in H1, intro, and at least twice in body): [keyword] Audience: [describe who will read this and their expertise level] Tone: [conversational / authoritative / technical / beginner-friendly] Structure: - H1 that includes the primary keyword - 120-word intro that names the reader's problem and promises a specific takeaway - 4–5 H2 sections - A concrete example in every section (not hypothetical) - A 100-word conclusion with a single clear action step Do not use filler phrases, generic advice, or conclusions that just restate the intro.

D
Analysis effort: medium

Data Analysis — Plain English

I have the following dataset. Analyse it and explain what it means in plain language — no jargon, no assuming statistical knowledge. Data: [paste your CSV, table, or key numbers] Answer these questions: 1. What is the single most important trend in this data? 2. What is the most surprising or counterintuitive finding? 3. What is the data NOT telling me that I might wrongly assume it does? 4. What one additional data point would change the interpretation most? 5. Summarise the whole thing in two sentences for someone who will not read the full analysis.

D
Coding effort: xhigh

Debugging — Root Cause Analysis

Use extended thinking to diagnose this bug. Walk through every possible cause systematically before settling on the most likely one. Error message: ``` [Paste error] ``` Relevant code: ``` [Paste code] ``` What I have already tried: [list attempts] Environment: [OS, runtime version, relevant packages and versions] Deliver: 1. Your hypothesis ranking (most → least likely cause, with reasoning) 2. The fix for the most likely cause with the specific line changes 3. The debugging steps I should take if the most likely fix doesn't work 4. How to prevent this class of bug in future

I
Writing effort: medium

Interview Prep — Q&A Coach

I have an interview for [job title] at [company type, e.g. 'Series B startup' or 'Fortune 500']. My background: [2-3 sentence summary of your experience] The role focuses on: [describe the key responsibilities] Generate: 1. The 8 most likely interview questions for this specific role 2. A strong answer framework for each (not a full script — the key points I should hit) 3. The 3 questions I should ask the interviewer to show genuine interest and strategic thinking 4. The one weakness I should avoid mentioning given this specific role and company type

P
Analysis effort: high

Product Requirements Doc (PRD)

Write a complete Product Requirements Document for the following feature: Feature name: [name] Problem it solves: [one sentence] Target user: [describe] Success metric: [how will we know it worked?] The PRD must include: 1. PROBLEM STATEMENT — 2 paragraphs: the current pain and why it matters now 2. PROPOSED SOLUTION — what we are building (not how) 3. USER STORIES — 5 stories in 'As a [user], I want [goal], so that [reason]' format 4. ACCEPTANCE CRITERIA — bullet list of testable conditions for each user story 5. OUT OF SCOPE — 3 things we are explicitly NOT building in v1 6. OPEN QUESTIONS — 3 decisions that are not yet made and who owns each 7. TIMELINE — rough phases (discovery / build / launch) with no dates, just sequence

Claude Sonnet 5 vs Other Leading AI Models (2026)

How Sonnet 5 compares to its stablemates and top competitors across the key dimensions that matter for everyday use.

Model Best For Context Reasoning Price (input/output)
Claude Sonnet 5 ✓ Everyday power use, coding, writing, agentic tasks 1M tokens Low / Medium / High / xhigh $2 / $10 per MTok
Claude Fable 5 Hardest frontier tasks, 200-page docs, full codebases 1M tokens Always-on adaptive thinking $10 / $50 per MTok
Claude Opus 4.8 Premium general use, long creative writing, research 1M tokens Standard (no effort selector) $7.50 / $37.50 per MTok
GPT-5.5 Broad general use, image understanding, plugin ecosystem 128K tokens Standard (no effort selector) $5 / $20 per MTok
Gemini 4 Multimodal tasks, Google ecosystem integration 1M tokens Thinking mode (on/off) $7 / $28 per MTok

Claude Sonnet 5 Prompting Tips

Do This:

  • Specify the reasoning effort level at the top of your prompt — it gives Claude an explicit directive
  • For coding tasks, include language, framework, and version — Sonnet 5 uses them to match conventions
  • Ask for structured output (numbered lists, sections) — Sonnet 5 formats it precisely
  • Provide rich context — the 1M context window means you rarely need to trim your documents
  • For agentic tasks, say "plan and execute" and ask Claude to narrate each step before doing it
  • End prompts with "explain any non-obvious decisions" to get reasoning surfaced inline

Avoid This:

  • Using xhigh effort for simple tasks — it's slower and adds no value for quick rewrites
  • Using low effort for multi-step reasoning tasks — the output will be shallow
  • Vague instructions like "make it better" — be specific about what dimension to improve
  • Over-constraining the format before stating the goal — goal first, format constraints after
  • Asking Sonnet 5 to predict real-time data — it has a knowledge cutoff, not live web access
  • Ignoring the model ID in API calls — always specify claude-sonnet-5 explicitly

Claude Sonnet 5 — Frequently Asked Questions

What is Claude Sonnet 5?

Claude Sonnet 5 (model ID: claude-sonnet-5) is Anthropic's newest general-purpose AI model, released June 30, 2026. It replaced Claude Sonnet 4.6 as the default model on Claude.ai and in GitHub Copilot. Sonnet 5 delivers near-Opus 4.8 performance at Sonnet pricing ($2 per million input tokens, $10 per million output tokens) with a 1M token context window and selectable reasoning effort — users can dial reasoning intensity from low (faster, cheaper) to xhigh (deep extended thinking) depending on the task.

What is the reasoning effort selector in Claude Sonnet 5?

Claude Sonnet 5 introduces a selectable reasoning effort level that controls how much thinking the model does before responding. The four levels are: low (fast, minimal reasoning — good for simple rewrites and lookups), medium (balanced — good for writing, analysis, and standard coding), high (deeper reasoning — good for research synthesis, complex debugging, strategic decisions), and xhigh (extended thinking — Claude Sonnet 5's maximum reasoning depth, for hard logic problems, mathematical proofs, and complex architectural decisions). You can specify the effort level directly in your prompt, as shown in the examples above.

How is Claude Sonnet 5 different from Claude Sonnet 4.6?

Claude Sonnet 5 is a major step up from Sonnet 4.6. The key improvements are: selectable reasoning effort (Sonnet 4.6 had no reasoning modes); near-Opus 4.8 performance on coding and analysis benchmarks; the same $2/$10 pricing as Sonnet 4.6; stronger agentic performance (multi-step tasks, tool use, long-horizon planning); and more consistent output formatting for structured tasks. Claude Sonnet 4.6 is no longer the default model on Claude.ai as of June 30, 2026.

Can I use these prompts for free?

Yes — all 20 prompts on this page are free to copy and use directly. Claude Sonnet 5 is available on the Claude.ai free tier (with usage limits), the Claude.ai Pro subscription, and via the Anthropic API at $2/$10 per million input/output tokens. Copy any prompt, paste it into Claude.ai or your API call, and add your specific context where indicated.

Which tasks is Claude Sonnet 5 best at?

Claude Sonnet 5 excels at: coding and debugging (especially with high or xhigh reasoning effort), research synthesis and analysis, long-form writing and content creation, agentic multi-step workflows, structured document generation (PRDs, API docs, briefs), and complex decision frameworks. For the hardest frontier tasks — 200-page legal contracts, full codebase architecture audits across millions of lines, and extremely long creative writing — Claude Fable 5 or Claude Opus 4.8 are better choices, though at significantly higher cost.

What is Claude Sonnet 5's context window?

Claude Sonnet 5 has a 1M token context window, the same as Claude Fable 5. This is enough to fit approximately 750,000 words, several full-length novels, or a large codebase with hundreds of files. For most real-world tasks — a long research paper, a full project codebase, or a book-length document — the context window is not a practical constraint. The maximum output length is 64K tokens per response.

More AI Model Prompt Generators