MonDive#24: GPT-5 vs Claude 4.5 Sonnet, Who Wins?

Welcome to the MonDive

Today in MonDive, we put two new AI titans head-to-head: OpenAI’s GPT-5 vs Anthropic’s Claude 4.5 Sonnet.

Same prompt, two models. The goal?
Find out who delivers better reasoning, creativity, and control — in real-world creator and business workflows.

Let’s dive in 👇

Want clarity on sales tax automation? Take the Avalara AvaTax product tour.

Sales tax is complex, but evaluating software shouldn't be. Avalara’s self-guided tour of AvaTax shows how to calculate tax in real time, track changing nexus thresholds, and apply the correct rates across states and jurisdictions. 
 
Used by more than 43,000 companies, AvaTax connects to platforms like NetSuite, QuickBooks, and Shopify, helping you automate compliance across more than 1,400 systems. You’ll see how to review flagged transactions by jurisdiction, create audit-ready reports, and improve accuracy without needing to meet with sales. 
 
If you’re ready to explore automation but short on time, this tour is designed for you.

🧠 Why This Matters

Not all next-gen AIs think the same.
Some are better at structured reasoning (logic, coding, precision).
Others shine in expression and tone (storytelling, empathy, nuance).

Here’s how they differ:

• GPT-5: Multimodal powerhouse — writes, codes, designs, and analyzes seamlessly.

• Claude 4.5 Sonnet: Context king — longer memory, refined prose, subtle emotional understanding.

For creators, marketers, and founders, knowing which one to trust for what could save hours every week.

Select ChatGPT-5 and Claude 4.5 Sonnet

Visit ChatGPT (OpenAI)

Go to ChatGPT

Log in with your OpenAI account.

In the top bar, click the model selector → choose GPT-5 (latest model).

Visit Claude (Anthropic)

Go to claude.ai 

Sign in using your Anthropic account (Google or email login).

Once inside, click the Model dropdown → select Claude 4.5 Sonnet.

1. Research & Multi-Source Synthesis

Input Example:

“Compare the 2024 EU AI Act, U.S. AI Bill of Rights, and China’s Generative AI Regulations. Create a summary table of similarities and differences, and conclude with policy implications for enterprise AI adoption.”

ChatGPT Result:

• Actively browses live government sources (EU Parliament, White House, CAC).

• Builds a formatted table with hyperlinks.

• Adds a short executive summary fit for presentation decks.

Claude Result:

• Reads and reasons through all three frameworks with a balanced narrative tone.

• Generates a clean comparison with legal nuance — e.g., ‘EU Act enforces accountability via risk tiers, while U.S. focuses on civil rights protection.’

• Maintains citations (if provided in context) but doesn’t browse.

👉 Winner: ChatGPT — real-time research edge and business formatting.

2. Analytical Reasoning

Input Example:

“Explain the impact of compound interest on a $10,000 investment over 10 years at 6% annual rate, compounded monthly. Include a short summary for beginners.”

ChatGPT-5 Result:

• Provides exact formula, breakdown, and step-by-step math.

• Includes a friendly summary: “Your money doubles in roughly 12 years at 6%.”

• Visualized with a short compound-curve example table.

Claude 4.5 Sonnet Result:

• Uses strong analogies (“your savings grow like a tree adding rings each year”).

• Simplifies math beautifully but skips exact monthly compounding formula.

• Feels more narrative than numeric.

👉 Winner: ChatGPT-5 (Precision + Math Accuracy)

3. CV / Resume Creation

Input Prompt:

“Write a modern, ATS-friendly resume for a Marketing Manager with 5 years’ experience in digital campaigns, analytics, and brand growth. Include measurable results and concise bullet points.”

ChatGPT-5 Result:

• Generates a polished, keyword-optimized layout.

• Strong action verbs, quantifiable metrics (“Increased ROI by 42%”).

• ATS-friendly, easy to scan.

• Slightly formal tone — lacks personality.

Claude 4.5 Sonnet Result:

• More story-driven resume style (“Blending creativity with data-driven insight”).

• Beautiful phrasing, but sometimes too narrative for corporate ATS.

• Naturally adds soft skills and emotional tone.

👉 Winner: ChatGPT-5 (Precision & Professional Formatting)

4. Presentation / Pitch Deck Creation

Input:

“Create a 10-slide investor presentation for a startup called EcoFuel, summarizing the problem, solution, market, traction, and business model. Keep tone persuasive and visually clear.”

ChatGPT-5 Result:

• Creates clean, bullet-based slide structure.

• Logical flow: Problem → Solution → Market → Model → Ask.

• Suggests visuals (“use a green gradient background and upward trend arrows”).

• Feels corporate and crisp — ready for Canva or PowerPoint.

Claude 4.5 Sonnet Result:

• Delivers a narrative-driven pitch (“The world runs on fuel — we make it sustainable”).

• Emotional storytelling that sells the mission.

• Slide suggestions come with tone notes (“begin with urgency, end with hope”).

👉 Winner: Claude 4.5 Sonnet (Storytelling & Emotional Persuasion)

5. Interview Preparation

Input Prompt:

What’s the latest update about Apple’s Vision Pro sales this week? Summarize from 3 trusted sources.

ChatGPT-5 Result:

• Provides structured Q&A with STAR method examples.

• Covers technical and behavioral questions.

• Great for quick mock-interview sessions.

• Feels mechanical — less coaching style.

Claude 4.5 Sonnet Result:

• Conversational coaching tone (“Here’s how to tell your story authentically”).

• Helps shape personal examples with emotional resonance.

• Great at tone coaching and confidence framing.

• Slightly lighter on technical depth.

👉 Winner: Claude 4.5 Sonnet (Empathy & Narrative Coaching)

6. Career Roadmap Planning

Input Prompt:

“Design a 1-year roadmap to transition from Graphic Designer to Brand Strategist — include learning paths, portfolio updates, and mindset shifts.”

ChatGPT-5 Result:

• Builds a clear 12-month timeline with quarterly milestones.

• Lists specific skills (brand architecture, positioning, analytics).

• Suggests courses, books, and measurable projects to track progress.

• Recommends portfolio upgrades (add brand-case studies, strategic briefs).

• Excellent structure but feels procedural — more plan than passion.

Claude 4.5 Sonnet Result:

• Frames the transition as a creative evolution, not just a career move.

• Divides the year into “seasons of growth” (Unlearn → Reframe → Build → Lead).

• Adds reflective prompts (“What story does your work tell about value, not visuals?”).

• Focuses on mindset and identity shift — from designer to storyteller of meaning.

• Inspiring and human, but less detailed on tools and metrics.

👉 Winner: Draw (Different Strengths)

Which AI helps you research and summarize information more effectively?

🥊 Results

ChatGPT-5 (OpenAI)

✅ Strengths:

• Exceptionally structured and multimodal — handles text, image, code, and data seamlessly.

• Generates clear, presentation-ready outputs (tables, timelines, frameworks).

• Excels at step-by-step logic, research synthesis, and measurable planning.

• Integrated tools for image creation, document analysis, and code reasoning.

• Ideal for professionals, analysts, and creators who need precision and iteration.

❌ Weaknesses:

• Can sound too formal or robotic for creative storytelling.

• Sometimes over-structures tasks that need emotional nuance.

• May require prompt tweaking to inject warmth or brand voice.

• Focused more on getting it right than making it feel human.

👉 Verdict:
Best when you need clarity, structure, and execution power — perfect for strategy documents, presentations, and research deliverables.
A dependable “engine room” for anyone who builds, plans, or analyzes.

Claude 4.5 Sonnet (Anthropic)

✅ Strengths:

• Writes with natural warmth, rhythm, and storytelling depth.

• Maintains context across very long conversations or documents.

• Excellent at tone coaching, brand voice, and creative reframing.

• Feels like a mentor or editor, not a machine — great for reflection and narrative work.

• Perfect for marketers, writers, leaders, and educators who value emotional connection.

❌ Weaknesses:

• Less rigid structure — outputs may lack clear timelines or bullets.

• Doesn’t browse the live web (context must be supplied).

• Occasionally romanticizes or over-personalizes professional tasks.

• Formatting consistency weaker for data-heavy reports.

👉 Verdict:
Best when you want empathy, storytelling, and audience resonance.
Ideal for crafting speeches, thought-leadership posts, or brand-driven messaging.
A creative partner that helps your ideas sound alive.

Know someone who may be interested?