Compare

Claude vs ChatGPT

Claude is our pick for code, dashboards and anything visual; ChatGPT stays open for everyday questions, PRDs and prompt writing. Since January 2026 more of our spend has gone to Anthropic.

By Jordan Metzner & Sam Nadler · From our episodes · Updated October 2026

Watch: Episode 11, Coding the Beat: Strudel, Claude Dashboards & OpenAI’s $1B Play from 12:41

The short answer

Pick Claude if you want it to build something: code, a dashboard from a spreadsheet, an interactive visual. Pick ChatGPT for everyday questions, writing a PRD or a prompt for another tool, and personal projects that live in one long conversation.

That split has held since our earliest episodes. In July 2025 Jordan summed up the market as ChatGPT winning "for most things" and Claude "for code." What changed is how much we lean on Claude: by January 2026 Jordan said Jordan and Sam's company was shifting more and more spend to Anthropic and less to ChatGPT.

What we built with each

Claude

CFO dashboard from a 10-K. Jordan gave Claude Google's 57-page 10-K PDF and asked for a dashboard "as if I'm a CFO." It returned revenue, net income, segment expenses, free cash flow and risks in one view (Episode 11, at 10:55). We do the same with QuickBooks exports.
Cholesterol medication 3D visualizer: built only by chatting with Claude, going from a 2D model to 3D with autoplay demos and labels (Episode 13).
Claude-generated n8n podcast guest outreach workflow: one prompt produced the workflow JSON, which Jordan imported into n8n (Episode 17).
ScreenEval: Sam's recruiter coach runs "Opus 4.5 for everything" (Episode 29).
Kalshi prediction market trading bot: Claude built the algorithm (Episode 26).

ChatGPT

Public API trading bot: the OpenAI API makes the buy and sell calls with confidence scores (Episode 8).
Ryz Labs Video Studio: ChatGPT powers the "enhance with AI" prompt rewriter (Episode 1).
Sam's ATS and the AI start-of-care packet generator: Sam wrote the PRDs in ChatGPT, then built elsewhere.
Sam's nutrition project: a ChatGPT project that tracks calories and macros from photos and voice notes. "One ChatGPT conversation is like a mini application" (Episode 4, at 14:38).

Some builds used both. The Ryz Labs Autoblogger writes with Claude and ChatGPT, and the LLM Math Roaster pitted four models against each other with ChatGPT as the judge (Episode 22).

Head-to-head

Our one direct test was the 10-K. We ran the same PDF through both:

And we did, just for context, put it in, the same document in OpenAI. It wasn't nearly as good. It didn't provide the immediate visual. The summary was pretty good, but I would I would argue that it was too much. It was too much to read.

— Sam Nadler, Episode 11

	ChatGPT	Claude
Dashboards from files	Long summary; extra prompts for four graphs	Full dashboard on the first try
Interactive visuals	Jordan: "wouldn't come out nearly as good"	"A very good web rendering engine"
Coding (Aug 2025)	GPT-5 "not a better coding algorithm"	Still ahead, per Jordan
Everyday use	PRDs, prompt writing, life admin	Zach Rattner keeps it open all day
Business model	Exploring ads (Jan 2026)	Rejected ads; enterprise focus

Where each one fell short

ChatGPT. The GPT-5 launch frustrated Sam: the auto router gave him the quick answer when he wanted the thinking one, so he had to re-prompt (Episode 8, at 23:27). Jordan's verdict a week in:

I think I mean, it's disappointing. It doesn't it's not a better coding algorithm, and it didn't beat, you know, the Anthropic model, so seems like everyone's gotta get back to work.

— Jordan Metzner, Episode 8

The ads news in Episode 28 pushed Jordan further: he said he wants to invest his company's time with Anthropic, which seems to have "more allegiance" to paying enterprise customers. Guest Sagi Waitzman said ads in the chat "turned me off from from OpenAI." Jordan also admits half of what he does in ChatGPT is copying output into Google Docs (Episode 23).

Claude. Cost, and availability. Jordan noted that even as plans got cheaper, "we're spending more money using it." In June 2026 Jordan had about 36 hours with Anthropic's Fable 5 before it was pulled. Jordan said the models left, 4.8 and 5.5, were "not nearly as good" (Episode 49, at 10:33). In Episode 46 he called Fable "very slow" and said it "seems expensive."

What our guests use

Tristan Wilson (construction): about 15 to 20% of firms he talks to have OpenAI enterprise accounts, partly so staff can upload sensitive files more securely. Executives use ChatGPT, "maybe Claude," or Grok daily.
Anis Bennaceur, Attention: "we're an anthropic shop here." Claude's UI shows each person's remaining daily usage, which makes spend caps workable (Episode 45).
Jackson Reiter (commercial real estate): firms "all seem to love Claude" but mostly use it for basic automation while the work stays in Excel.
Zach Rattner: Claude is open from the moment he gets to his desk. Emanuele Melis uses ChatGPT to make custom coloring-book PDFs for his kids.

Which one we reach for now

July to September 2025: ChatGPT for general use, Claude for code and visuals.
October 2025: Jordan said that since GPT-5, OpenAI had "taken the lead on the chat bot for sure," with Codex leading in code (Episode 15).
January to February 2026: back to Anthropic. Jordan said "the Anthropic models are are the best way, no doubt" (Episode 29), and called Opus the best model on the market when 4.6 shipped (Episode 30).
October 2026: Jordan tested the new Opus and said it was "performing quite well and significantly cheaper than using Fable." He also uses Codex a lot and found the lower-level GPT-6 models cost effective (Episode 53).

So today: Claude for building and analysis, ChatGPT for quick everyday help, and OpenAI's Codex for much of Jordan's coding. Jordan's own rule is that loyalty follows quality: if another lab's models become "top notch," we would probably move. For the coding tools themselves, see Claude Code vs OpenAI Codex; for Google's model, see Claude vs Gemini and ChatGPT vs Gemini.

FAQ

Is Claude better than ChatGPT?

For coding and for turning files into visual dashboards, it was in our tests. When we gave both the same 57-page Google 10-K, Claude built a full dashboard right away while ChatGPT gave a long summary and needed more prompts for four graphs.

Is Claude better than ChatGPT for coding?

Jordan said in August 2025 that GPT-5 was not a better coding model and did not beat Anthropic's. OpenAI's Codex later took the lead for a while, so for coding tools see our Claude Code vs Codex comparison.

What is ChatGPT better at than Claude?

We use ChatGPT for PRDs, prompt writing, Suno song prompts and everyday life tasks like Sam's calorie tracking project. Jordan also finds its memory helpful.

Which do businesses use, ChatGPT or Claude?

Both. Tristan Wilson sees OpenAI enterprise accounts in about 15 to 20% of construction firms he talks to, while Attention calls itself an Anthropic shop and real estate firms told Jackson Reiter they love Claude.