Model comparisons - tested October 2026 by Vitus Pruser. How we test.
Verdict
GPT-6 Astra took 9 of 10 prompts on completeness and reliability. Perplexity Sonar Pro was the fastest model in the whole round (average 8 seconds) and the only one citing live sources, which won it the EU VAT research answer. But it also made the round's only real error, quoting 5.65% for an average that is 4.90%, and its ordering of China ahead of India on 2024 population follows the CIA World Factbook rather than the UN consensus. Pick Sonar for current-facts questions with receipts; pick GPT-6 Astra for everything else, and check Sonar's arithmetic.
Verdict by task
| Task | Winner | Why |
|---|
| Writing | GPT-6 Astra | Sonar was shortest but lightest: no subject line and fewer facts kept in the rewrite. GPT-6 Astra produced fuller, on-brief copy both times. |
| Coding | GPT-6 Astra | Sonar's slugify passed all its tests and is clean, but its rounding fix silently changes the rounding policy; GPT-6 Astra was correct on both, though it pulls in one dependency. |
| Research | Split | Sonar won the EU VAT question with real citation markers and complete figures; GPT-6 Astra won populations with exact UN estimates where Sonar's CIA source put China first. |
| Data analysis | GPT-6 Astra | Sonar mis-computed the average growth rate (5.65% instead of the real 4.90%), the only arithmetic slip in the round; GPT-6 Astra was correct on both tasks. |
| Reasoning | GPT-6 Astra | Both solved both puzzles; Sonar hedged with "one of the shortest" when its route is the shortest, and GPT-6 Astra presented state and check more clearly. |
Results, prompt by prompt
writing - winner: GPT-6 Astra
Prompt: Write a 120 word product update email to existing customers of a small accounting software company announcing that bank feeds now sync every hour instead of once a day. Plain, friendly, no exclamation marks, end with one clear call to action.
- Perplexity Sonar Pro - pass, no hallucinations, 13.7s. 91 words, readable, ends with a soft call to action. No subject line and lightest on detail.
- GPT-6 Astra - pass, no hallucinations, 16.1s. 124 words, plain and friendly, ends with a clear log in call to action. No subject line.
writing - winner: GPT-6 Astra
Prompt: Rewrite this paragraph so a 12 year old understands it, keep every fact, max 80 words: 'Compound interest is interest calculated on the initial principal and also on the accumulated interest of previous periods. Because the base grows each period, the growth is exponential rather than linear, which is why long time horizons matter far more than the interest rate alone.'
- Perplexity Sonar Pro - pass, no hallucinations, 7.4s. 55 words, keeps most facts but swaps exponential-versus-linear for the vaguer "faster and faster".
- GPT-6 Astra - pass, no hallucinations, 12.5s. 66 words, keeps every fact including exponential versus linear growth. Cleanest rewrite.
coding - winner: GPT-6 Astra
Prompt: Write a Python function slugify(title: str) -> str that lowercases, replaces any run of non alphanumeric characters with a single hyphen, strips leading and trailing hyphens, and transliterates common accented Latin letters (e, a, o, u with accents, and the German sharp s) to ASCII. Include 5 pytest test cases. No external dependencies.
- Perplexity Sonar Pro - pass, no hallucinations, 9s. All 5 tests pass, clean implementation handling ß and ẞ explicitly.
- GPT-6 Astra - pass, no hallucinations, 22.6s. Leanest correct solution: casefold() handles ß naturally. All 5 tests pass and it survives extra edge cases (Ü, Ñ).
coding - winner: GPT-6 Astra
Prompt: This JavaScript has a bug: users report the total is sometimes off by one cent. Find the bug, explain it in two sentences, and give the corrected function.
function total(items) {
let sum = 0;
for (const it of items) {
sum += it.price * it.qty * (1 + it.taxRate);
}
return Math.round(sum * 100) / 100;
}
- Perplexity Sonar Pro - partial, no hallucinations, 7.3s. Per-line cents fix helps accumulation but silently changes rounding policy and misses the boundary case.
- GPT-6 Astra - pass, no hallucinations, 17s. Correct diagnosis and a Decimal.js fix that is exact on every boundary case, but it adds an npm dependency.
research - winner: Perplexity Sonar Pro
Prompt: What is the current EU VAT registration threshold for small businesses under the EU wide SME scheme that started in 2025, in euros, and which regulation introduced it? Answer in 3 sentences and say if you are unsure about any figure.
- Perplexity Sonar Pro - pass, no hallucinations, 7.4s. Correct figures plus real citation markers and the €85,000 national cap. The most sourced answer.
- GPT-6 Astra - pass, no hallucinations, 13.4s. Correct figures (€100,000 union threshold, €85,000 national cap), correct instrument (Directive 2020/285), and honestly flags it cannot verify later amendments.
research - winner: GPT-6 Astra
Prompt: List the 5 largest countries by population in 2024 with their approximate population, and name the source you would cite for these figures. Table format.
- Perplexity Sonar Pro - pass, no hallucinations, 6.2s. Real citations, but orders China first on CIA World Factbook figures, against the UN consensus where India has led since 2023.
- GPT-6 Astra - pass, no hallucinations, 11s. Most precise figures (India 1.451b, China 1.419b), matching UN World Population Prospects 2024 mid-year estimates, with the source named.
data - winner: GPT-6 Astra
Prompt: Here are monthly revenue figures in EUR for a shop: Jan 41200, Feb 38900, Mar 45100, Apr 47800, May 46200, Jun 52300, Jul 55900, Aug 51400, Sep 58700. Compute the month over month growth for each month, the average growth rate, and the best and worst month. Show the numbers in a table and state the totals.
- Perplexity Sonar Pro - partial, 1 hallucination, 11s. Every monthly figure right, but its computed average growth of 5.65% is wrong: the real mean is 4.90%.
- GPT-6 Astra - pass, no hallucinations, 20.6s. Every month over month figure correct (avg 4.90%), plus euro changes, totals and a multi-metric best and worst table.
data - winner: GPT-6 Astra
Prompt: A SaaS company has 1,240 customers paying an average of 49 EUR per month and loses 3.1 percent of customers each month. New signups are 85 per month. Estimate the customer count and MRR after 12 months assuming these rates hold. Show your working and round to whole numbers.
- Perplexity Sonar Pro - pass, no hallucinations, 8.6s. 1,713 and 83,937 correct, rounds intermediates more loosely.
- GPT-6 Astra - pass, no hallucinations, 23s. Closed-form solution, 1,713 customers, MRR 83,920 computed on the unrounded count, which is the more rigorous route.
reasoning - winner: GPT-6 Astra
Prompt: I have a 3 litre jug and a 5 litre jug and unlimited water. Give the shortest sequence of steps to measure exactly 4 litres. Then say how many steps it took.
- Perplexity Sonar Pro - pass, no hallucinations, 7.5s. Optimal 6 steps; hedges with "one of the shortest" when it is the shortest.
- GPT-6 Astra - pass, no hallucinations, 11.1s. Optimal 6 steps with per-step jug states shown.
reasoning - winner: GPT-6 Astra
Prompt: Three colleagues need to split a 187 EUR dinner. Anna had no drinks and her food was 41 EUR. Ben and Clara shared two bottles of wine totalling 58 EUR and their food was 44 EUR each. Who owes what, so that everyone pays exactly for what they consumed? Show the check that the parts add up to 187.
- Perplexity Sonar Pro - pass, no hallucinations, 6.2s. 41, 73, 73 correct, clean per-person breakdown.
- GPT-6 Astra - pass, no hallucinations, 8.6s. 41, 73, 73, check to 187. Correct and cleanest.
Frequently asked questions
Is Perplexity or ChatGPT better?
For general work, ChatGPT: GPT-6 Astra won 9 of 10 blind-ranked prompts in our October 2026 round. Perplexity Sonar Pro is the better pick only when the question needs live sources and citations, and it was the fastest model we tested at about 8 seconds per answer.
Is Perplexity or ChatGPT better for research?
Sonar Pro won the EU VAT threshold answer with real citations and complete figures, but on the population table its CIA World Factbook source ordered China ahead of India, against the UN consensus. For current events with receipts pick Perplexity; for careful sourced facts we found GPT-6 Astra more consistent.
Is Perplexity or ChatGPT better for coding?
Sonar Pro wrote a clean slugify that passed all its own tests, but its cent-rounding fix changes the rounding policy without saying so. GPT-6 Astra was correct on both coding tasks. For code you rely on, GPT-6 Astra.
Which is cheaper, Perplexity or ChatGPT?
Perplexity Sonar Pro. Across the ten prompts Sonar cost $0.09 in total against $0.21 for GPT-6 Astra, and it was the fastest model in the round.
Can I use Perplexity and ChatGPT in the same place?
Yes. Krater gives you Perplexity Sonar Pro, GPT-6 Astra and 500+ other models in one workspace, and the Model Arena sends one prompt to several models side by side, which is how this comparison was run.
How was this comparison tested?
All five models (GPT-6 Astra, Claude Fable 5.1, Gemini 3.1 Pro Preview, Grok 4.3 and Perplexity Sonar Pro) received the same ten prompts side by side in the Krater Model Arena on 7 October 2026, with default settings and no system prompt; this page compares the two you are looking at. We recorded time to a complete answer and the real model cost per answer, ran the generated code through its own test suites, fact checked every claim, scored completion as pass, partial or fail, and then ranked the answers blind. The full procedure is on our methodology page.
Will this comparison be updated?
Yes. Flagship comparisons are rerun monthly and whenever a provider ships a new version. The tested month and model versions are shown at the top of the page.
