Small business▲+2Since September 2026: 0% → 2%, +2 points. Inside the 11-point floor: within noise. Read over the models both editions asked.
2%from 0% in September 2026
Inside the floor by 9 pointsRank 62 → 16 of 131
1 category · Named, not rankedPyTorch
20Judge labels
3First choices
3Negative labels
10 / 14Models named it
1Category
October 2026 Edition. Every number here is derived from the raw labels under vendor table v2026-10.8, every buyer segment counted.
Standing
9 labels, too few to rank
A product needs 10 labels in a category before a share or quadrant is stated. PyTorch was named 9 times in ML platforms, where Azure Machine Learning led with 21%. The labels and the evidence are below, counted exactly.
By buyer segmentRead the same way at every buyer size.
Small business16 of 139
2%8 labels
under the label threshold
Mid-market9 of 107
4%9 labels
under the label threshold
Enterprise60 of 65
0%3 labels
under the label threshold
In ml platforms · each standing computed within its segment · bars are 0 to 100 · the accent bar is the product's own best reading
Standing by category
Every category where a model named PyTorch for a mid-market B2B company. Share is first choices across the direct, paraphrase, budget and scale prompts; rank is within every product named in that category.
| Category | Function | Share | Rank | Negative rate | Labels | Quadrant | Since September 2026 |
|---|---|---|---|---|---|---|---|
| ML platforms | Data platform | 4% | 9 of 107 | 0% | 9 | under 10 labels · led by Azure Machine Learning at 21% | ▲+4Since September 2026: 0% → 4%, +4 points. Inside the 11-point floor: within noise. Read over the models both editions asked. |
Movement
Since September 2026, PyTorch held within the floor in its one category: no change cleared 11 points.
Best category · ML platforms · by segment
Mid-market▲+4Since September 2026: 0% → 4%, +4 points. Inside the 11-point floor: within noise. Read over the models both editions asked.
4%from 0% in September 2026
Inside the floor by 7 pointsRank 18 → 9 of 97
EnterprisenewNew since September 2026: not ranked then, 0% now.
0%from 0% in September 2026
Inside the floor by 11 pointsRank 50 of 55, unchanged
All categories · Mid-market
| Category | September 2026 | Now | Change | Reading | Rank |
|---|---|---|---|---|---|
| ML platforms | 0% | 4% | ▲+4Since September 2026: 0% → 4%, +4 points. Inside the 11-point floor: within noise. Read over the models both editions asked. | Within noise | Rank 18 → 9 of 107 |
Shares here are read over the models both editions asked, so they can differ by a point or two from the standing above, which counts every model in this edition.
The floor is 11 points of share, measured: how far the models move a leader on their own when the same questions are asked twice with nothing changed. A larger change is movement; a smaller one is noise, and both are shown. Movement is read over the twelve models both editions asked; GPT-6 Luna, Muse Glimmer 30B joined this edition and are in the standing but not yet in the comparison. How the floor is measured · The editions
By model
How each model treated PyTorch across every prompt where it was named for a mid-market B2B company. Fourteen models, six prompts per category.ShowHide
By model
How each model treated PyTorch across every prompt where it was named for a mid-market B2B company. Fourteen models, six prompts per category.
ShowHide| Model | First choice | Alternative | Mention | Negative | Labels |
|---|---|---|---|---|---|
| Claude Haiku 4.5 | 1 | 0 | 0 | 0 | 1 |
| GPT-5.4 mini | 0 | 0 | 0 | 0 | 0 |
| Gemini 3.5 Flash | 0 | 0 | 0 | 0 | 0 |
| Perplexity Sonar | 0 | 0 | 0 | 0 | 0 |
| Grok 4.20 | 0 | 0 | 0 | 0 | 0 |
| Mistral Small | 0 | 0 | 1 | 0 | 1 |
| DeepSeek V4 Flash | 0 | 1 | 0 | 0 | 1 |
| Llama 4 Maverick | 0 | 0 | 0 | 0 | 0 |
| Qwen 3.7 Flash | 0 | 1 | 1 | 0 | 2 |
| Kimi K2 | 1 | 1 | 0 | 0 | 2 |
| GLM 4.7 FlashX | 0 | 0 | 0 | 0 | 0 |
| MiniMax M2.5 | 0 | 1 | 0 | 0 | 1 |
| GPT-6 Luna | 0 | 0 | 0 | 0 | 0 |
| Muse Glimmer 30B | 0 | 0 | 1 | 0 | 1 |
By framing
Which of the six questions produced the naming. By model says how often; this says asked what. The first-choice count on the right carries the marks of the models that produced it.
FramingLabels by classFirst choices
Direct2 labelsNone
Paraphrase0 labelsNone
Comparative3 labelsNone
Budget-constrained9 labels3
Scale-constrained1 labelNone
Negative5 labelsNone
First choiceAlternativeMentionNegative20 labels in all, every segment counted; 3 of the 3 first choices count toward share, since the comparative and negative framings do not. The bar is one segment per label class, to scale within the framing.
What the models said for it
Verbatim evidence the judge attached to positive labels.
“Start with open-source tools like Scikit-learn for traditional ML or PyTorch for deep learning” Claude Haiku 4.5 · ML platforms · budget prompt · first choice
“start with open-source tools like TensorFlow or PyTorch combined with Google Colab” Kimi K2 · ML platforms · budget prompt · first choice
“Consider open-source options (TensorFlow, PyTorch) with proper security management” MiniMax M2.5 · ML platforms · negative prompt · alternative
“Open-source frameworks (scikit-learn, PyTorch, TensorFlow, Hugging Face)” Kimi K2 · ML platforms · negative prompt · alternative
And against it
Verbatim evidence attached to negative labels. A warning on a product with few labels is a warning; on a product with many, it is one voice among them.
No model argued against it.
Named alongside
The products named in the same answers as PyTorch, over the 20 answers that named it. Took the first choice instead counts the answers where the other product was the first choice and PyTorch was named but was not.ShowHide
Named alongside
The products named in the same answers as PyTorch, over the 20 answers that named it. Took the first choice instead counts the answers where the other product was the first choice and PyTorch was named but was not.
ShowHideProductSame answerTook the first choice insteadHead to head
A head-to-head page exists where both products are among a category's top eight. The other rows are the same fact without a page behind them, so they link to the product instead.
What carried it into the answer
The sites and pages cited by the answers that named PyTorch. A fact about retrieval, not a lever on the model.
Citations exist only for the models that return a source list, five of the fourteen in this edition, so these counts come from 17 of the 20 answers that named PyTorch and are not a share of its labels.
Domains cited
g2.com12
deepusecase.com10
byteplus.com9
spotsaas.com7
checkthat.ai6
eweek.com6
it-ai-index.com6
dataiku.com5
learn.g2.com5
us.fitgap.com5
Seventy-one of the seventy-one domain citations in answers naming PyTorch came from somebody else's page.
Pages cited
Pages are listed as the models cited them.
Search and answers
pytorch.org is not in Google's top ten for any of the category's searches. In the answers, PyTorch takes 4% of first choices and Azure Machine Learning takes 21%.ShowHide
Search and answers
pytorch.org is not in Google's top ten for any of the category's searches. In the answers, PyTorch takes 4% of first choices and Azure Machine Learning takes 21%.
ShowHideIn search
Google, US estimatesPosition for “machine learning platform”
–
not in the top ten
Position for “machine learning platforms”
–
not in the top ten
- Searches for its name, Google
- 40,500 a month (“pytorch”)
- AI search demand for its name, est.
- 4,030 a month
- Organic visits to its site
- about 150,980 a month
- Searches its site ranks for
- 10,786 · 3,080 in the top three
- Sites linking to it
- 28,487
- Paid Google search
- No ads found in the estimate; that does not mean it runs none
Ads on Google
None found in Google's ad transparency records for the US
In answers
This editionShare of first choices
4%
rank 9 of 107 in ML platforms
Segment leader
21%
Azure Machine Learning
- First choices
- 3 across its categories
- Named in
- 20 answers
- Its own site cited
- in 17 of the answers that named it
Search figures are US estimates from DataForSEO, read September 28, 2026; AI search demand is its modeled, directional estimate, not a count of queries to any assistant. The answers are this edition's. Two measurements side by side: neither is read as the cause of the other.
What it publishes
ShowHide
What it publishes
ShowHideAddresses on pytorch.org
10,522
subdomains included
Content
483
counted in the table below
Documentation
3,940
Product · Integration
1 · 0
| Kind | Pages | Last 90 days | 2025-10 to 2026-09 | Latest | Categories named |
|---|---|---|---|---|---|
| Blog | 276 | 24 | 2026-09-23 | ||
| Comparison | 75 | 1 | 2026-07-30 | ||
| Glossary or explainer | 53 | 0 | 2019-01-16 | ||
| Conference or event | 30 | 22 | 2026-12-10 | ||
| News or press | 22 | 3 | 2026-09-16 | ||
| Report or ebook | 16 | 0 | 2023-01-01 | ||
| Template or tool | 9 | 0 | 2023-01-01 | ||
| Case study | 2 | 0 | 2023-01-01 |
Most recent
- 2026-12-10Pytorch day japan 2026 conference or event
- 2026-12-06Neurips 2026 conference or event
- 2026-11-30Nvidia gtc washington dc conference or event
- 2026-11-21Pytorch day korea 2026 conference or event
- 2026-11-06Since ai hackathon 2026 conference or event
- 2026-10-20Pytorch conference 2026 conference or event
Events in person
Upcoming
- November 6, 2026 to November 8, 2026Since AI Hackathon 2026Turku, Finland, Turku, Finland
- November 21, 2026PyTorch Day Korea 2026Seoul, Korea, Seoul, Korea, Republic of · hosts
- December 10, 2026PyTorch Day Japan 2026Tokyo, Japan · hosts
As the event pages on pytorch.org state them, read September 28, 2026.
How it is countedHide how it is counted
Every page pytorch.org exposes, subdomains included. Kind is read from the address and title. The last 90 days, the latest date and the twelve months count pages by when they were published, from the site's feeds, a date in the address, or the page's own publication date, read from up to a hundred of its most recently changed pages; a page that says only when it last changed is counted in its kind but not in when, so the recent counts are a floor, and a kind none of whose pages gives a publication date reads undated. An event counts as online when its address or title says so (webinar, on demand, virtual or online summit); a conference, summit, trade show, expo or roadshow that does not say so is counted as a conference or event, which on a vendor's site is mostly in person. Read September 29, 2026.
Names read as PyTorch
What the judge wrote, as written, with how often. The vendor table decides that these count as PyTorch; a claim can dispute any of them.
PyTorch (Meta) 1
Follow PyTorch
An email the morning each edition publishes: where this product moved, where it held, and by how much against the noise floor. One address, confirmed by a click; a stop link in every email.
Already following? Everything you follow, with a stop for each.
The company
PyTorch is its own company.ShowHide
The company
PyTorch is its own company.
ShowHideIn its own words
Stated by the vendor, not checked- Positioning
- PyTorch Foundation is the deep learning community home for the open source PyTorch framework and ecosystem.
- Certifications
- PyTorch Certification
- Customers named
- Amazon AdvertisingSalesforceStanford UniversityShopifyLinkedInIBM ResearchOpenSynthIntel
- Not stated on the pages read
- Who it is for, price, starting price, free plan or trial, integrations, hosting
What PyTorch's own pages state, read September 28, 2026: pytorch.org, pytorch.org/features, pytorch.org/case-studies. A claimed page can correct any of them.
Is this your product?
Claim this page
Claiming is free and changes nothing in the data. A claimed page shows a verified contact who is told when each edition publishes and when PyTorch's standing changes by more than the noise floor; the right to propose corrections to the vendor table, meaning names the judge wrote that should or should not read as PyTorch, applied by version and listed in the change log; and a one-line description supplied by the vendor and marked as such.
What a new claim receivesHide what a new claim receives
A new claim receives the current edition's vendor brief for PyTorch by email, built from the raw record of the edition. It shows:
- where PyTorch is named, by buyer and by framing, and which cells hold its first choices;
- the claims the models make when they name it, ranked, with the strongest and the weakest quoted;
- its vocabulary against the segment leader's, and the pages the models cited;
- who was chosen in the answers that did not name PyTorch, and every reason the record gives;
- a battlecard for each top rival: the head-to-head split, why they win, and the reservation quoted against them;
- one page of published figures cleared to show a buyer.