Is DeepSeek Better Than ChatGPT’s Models? DeepSeek V4.1 vs GPT-6.1 Sol

Not the apps, the models: the same board deck, the same data and the same pipeline, with only the model changed. One cost seven times as much as the other, and found less.

By Ed · Published November 5, 2026 · 4 min read

Twenty-five years starting companies, consulting, owning a multi-million-dollar P&L and leading business intelligence for large enterprises.

On the same board deck, DeepSeek V4.1 Flash cost 30 cents and GPT-6.1 Sol $2.08. DeepSeek went deeper: it found Meta prospecting losing money by June, and made sharper recommendations. Sol was faster and the most careful with its labels. DeepSeek found 5 of 8 story points in our checklist, Sol 3. We tested the models, not the ChatGPT app.

How this was tested. One prompt, “Northvale 2026 year-to-date board review, with an executive summary and recommendations”, run once per model in the Quiriz PowerPoint add-in on November 5, 2026. Northvale is a fictional US apparel company with stories planted in its data: Google and Meta ad exports and a HubSpot wholesale pipeline, through September 15, 2026. Every step ran on the model under test: plan, build (answering each slide’s question from the data), gap analysis, write and critic; see how a deck is built. Three small parts stayed fixed for every run (chart styling, question routing and a grounding check). Cost is what OpenRouter billed. Build time runs from pressing Build to the last slide written, and leaves out the plan, which waits for a person to approve it. One run per model, so treat close results as a tie. Full method on the benchmark page.

At a glance

	DeepSeek V4.1 Flash	GPT-6.1 Sol
Billed cost	$0.30	$2.08
Build time	8.2 min	5.4 min
Analysis calls	67	33
Content slides	14	11
Stories found	5 of 8	3 of 8
Meta prospecting	Below 1.00 in June, $385 lost	Cost up 50%, conversions flat
Southeast	Fewest wins	Not covered
Price as a loss driver	Not singled out	The top loss pool overall
Recommendations	Cut prospecting, fund Brand-Search, shift to retargeting, re-point the Southeast	Pricing, buyer commitments, a NonBrand test

Where DeepSeek won

It found the problem. Both models saw Prospecting-Lookalike take most of the added Meta spend. Sol stopped at “cost up 50%, conversions flat”. DeepSeek followed the return month by month and drilled into June, where both Lookalike ad sets returned less than they cost. That turns “less efficient” into “losing money”, and its first recommendation, cut it in October, follows. It also put sales coverage on the Southeast, where Sol saw nothing to act on.

DeepSeek V4.1 Flash deck: both June prospecting ad sets ran below 1.00 — DeepSeek’s June drill-down: both Lookalike ad sets below 1.00.

Where Sol won

Speed and care. Sol finished in 5.4 minutes against 8.2, labeled every ad figure as platform-reported, said what the data could not answer, and ranked loss reasons year over year, which showed pricing as the largest pool of lost value.

GPT-6.1 Sol deck: ranked priorities: pricing, buyer commitments, then budget reallocation — GPT-6.1 Sol’s ranked recommendations.

Why the price gap is 7x, not 13x

Sol’s list price is about 13 times DeepSeek’s per input token. DeepSeek took 67 analysis calls to Sol’s 33 and read twice the tokens, and caching discounted both bills. More in what an AI board deck really costs.

Verdict

For a monthly deck, DeepSeek V4.1 Flash is the better buy: it finds more, for less. If speed and cautious wording matter most, Sol is the pick. On this test, Claude Opus 5.5 was the only model that found the ad story, the Southeast story and the price behind it.

Run the same deck on your own data

Describe the deck in PowerPoint and see the source of every figure. Free to start.

Try Quiriz free →

Frequently asked questions

Is DeepSeek better than ChatGPT?

We did not test the ChatGPT app. Against OpenAI’s GPT-6.1 Sol model on the same board deck, DeepSeek V4.1 Flash was 7 times cheaper, found more of the stories in the data (5 of 8 against 3) and made sharper recommendations; Sol was faster and more careful with its labels.

How much cheaper is DeepSeek than GPT-6.1 Sol?

About 7 times per deck in our test: 30 cents against $2.08. List prices differ by more (15 cents against $2 per million input tokens); Sol’s fewer analysis steps narrowed the gap.

Which went deeper, DeepSeek or GPT-6.1 Sol?

DeepSeek: it followed Meta returns month by month and found a campaign losing money in June, where Sol described the same campaign as less efficient. Sol was the more cautious of the two, labeling every ad figure as platform-reported.

Which is faster?

GPT-6.1 Sol: 5.4 minutes against DeepSeek’s 8.2.