Something unusual happened this week. On February 5, 2026, both OpenAI and Anthropic released their flagship AI models within hours of each other. GPT-5.3 Codex and Claude Opus 4.6 arrived almost simultaneously.

This wasn’t a coincidence. This was competition at its finest.

But here’s what really matters: both companies are solving different problems. And understanding which one fits your work can save you hours every week.

One Topic: GPT-5.3 Codex and Claude Opus 4.6 Release

The Shift from Chat to Work

I’ve been using AI tools daily for over a 3 years now. The biggest change I’m seeing isn’t about smarter answers. It’s about AI doing actual work.

These new models don’t just respond to questions. They can operate your computer, manage files, write code across multiple files, and work on projects for hours without losing track.

Think of it this way: earlier AI was like having a really smart friend you could ask questions. Now it’s like having a colleague who can sit at your desk and finish projects while you focus on other things.

• GPT-5.3 Codex is built for execution. It’s 25% faster than its predecessor and excels at terminal work, coding, and operating desktop applications. If you need something built quickly, this is your tool.
• Claude Opus 4.6 is built for thinking. It can hold 1 million tokens in memory (roughly 750,000 words) and excels at research, planning, and working through complex documents without forgetting context.

What This Means for Your Daily Work

Here’s where it gets practical.

I recently used Claude Opus 4.6 to analyse a 250-page report. I asked it to find specific patterns, compare sections, and pull out contradictions. It held the entire document in memory and gave me insights I would have missed reading manually. That’s the power of its massive context window. Earlier, Google Gemini had long context but Claude reasoning and voice is much better.

The same time, I used GPT-5.3 Codex to automate a repetitive task that involved moving files, renaming them based on specific patterns, and updating tags. It handled the entire workflow through the terminal in minutes.

Two different tools. Two different strengths.

Use GPT-5.3 Codex when you need:

• Fast execution on coding tasks
• Terminal and command-line automation
• Building applications or websites quickly
• Desktop automation and file management

Use Claude Opus 4.6 when you need:

• Deep analysis of large documents
• Research across multiple sources
• Long-form planning and strategy work
• Working with massive codebases that need careful review

The Free Access Window (Don’t Miss This)

Here’s the important part: OpenAI is making Codex free for ChatGPT Free and Go users for a limited time. Plus and Pro users get double the rate limits for the next two months.

This is your chance to experiment without commitment.

I recommend picking one repetitive task you do every week. Could be organising files, drafting reports, analysing data, or building a small tool. Give these models that task and see what happens.

You won’t master them by reading about them. You’ll master them by using them.

How I’m Using Both

I’ve settled into a pattern that works for me:

1. Research phase: Claude Opus 4.6 for reading industry reports, analysing competitive data, planning, and writing PRD.
2. Execution phase: GPT-5.3 Codex for building tools, automating workflows, and handling technical tasks using Codex Mac App or VS Code.
3. Review phase: Back to Claude for checking quality, finding gaps, and strategic thinking.

It’s not about choosing one over the other. It’s about using the right tool for the right job.

The Real Breakthrough

What excites me most isn’t the raw power of these models. It’s what they free you to do. I’m spending less time on repetitive tasks and more time on work that actually requires human judgment. Strategy. Relationships. Creative thinking. The stuff AI can’t replicate yet. That’s the shift happening right now. AI isn’t replacing knowledge work. It’s removing the friction so we can focus on what matters most.

Two models released on the same day. Both pushing boundaries. Both solving real problems. The question isn’t which one is better. The question is: which one will you use to get back 10 hours this week?

Quick tip: Don’t wait for the perfect use case. Start with something small and annoying. That’s where you’ll learn fastest.

Interested in travel or photography, read last week’s LensLetter newsletter about Stop Shooting and Start Telling Stories.

Read last week’s JustDraft about OpenClaw.

Two Quotes to Inspire

The best AI tool isn’t the smartest one. It’s the one that removes friction from your most repetitive work, so you can focus on what actually requires your brain.

Strategy isn’t about having all the tools. It’s about knowing exactly which tool solves which problem – and having the discipline to use the right one at the right time.

One Passage From My Bookshelf

The kernel of a strategy contains three elements: a diagnosis, a guiding policy, and coherent action. The diagnosis defines or explains the nature of the challenge. A good diagnosis simplifies the often overwhelming complexity of reality by identifying certain aspects of the situation as critical. A good guiding policy outlines an overall approach for overcoming the obstacles identified in the diagnosis. It channels action in certain directions without defining exactly what shall be done.

Coherent actions are feasible coordinated policies, resource commitments, and actions designed to carry out the guiding policy. Many strategies fail because they are simply not coherent. The most typical source of incoherence is the application of effort to too many different problems or objectives at once. A good strategy doesn’t just draw on existing strength; it creates strength through the coherence of its design. Most complex organizations spread themselves thin, as if every unit’s strategy were to serve every possible constituency. By contrast, good strategy makes a choice about which areas will receive the most resources and attention – and which will not.

📚From “Good Strategy Bad Strategy” by Richard Rumelt