“We don’t know how big it is; we don’t know what the architecture is, we don’t know how much energy was used; we don’t how many processors were used; we don’t know what it was trained on etc.”
Our reading, the sentence that was checked:
OpenAI has not disclosed GPT-4's size, architecture, energy use, processor count, or training data.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, GPT-4’s successes, and GPT-4’s failures, paragraph 8.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-gpt-4s-successes-and-gpt-4s-failures-17c3ce-a2
We recorded 12 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-29. The source may have been edited after 2023-03-15. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
From the speaker
- We don’t know how big it is
- we don’t know what the architecture is
- we don’t know how much energy was used
- we don’t how many processors were used
- we don’t know what it was trained on
Our interpretation
- We means the public research community, not OpenAI's own employees.
- The plain reading concerns GPT-4's specific technical specifications: parameter count, detailed architecture, training energy consumption, processor count, and identifiable training-data composition.
- A general model-family label or broad description of data sources does not disclose those specifications.
- The target is the disclosure record at launch and publication, March 14–15, 2023. Later disclosures cannot change what was public then.
Counts as supported if: The release materials and contemporaneous statements establish that the listed specific technical details were withheld from the public.
Counts as contradicted if: Publicly available materials by publication disclose the listed specifications. Disclosure of only some would qualify the claim rather than refute all of it.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Supported
The dated evidence matches what the claim committed to, under the criteria frozen before grading.
Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.
- Evidence as of
- Deadline state when assessed
- No horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
On March 15, 2023, OpenAI's chief scientist confirmed that OpenAI did not disclose GPT-4's parameter count or training data [e1]. The original technical report described GPT-4 as Transformer-based, providing a general architecture label [e2]. These disclosures support the reading that specific technical details were withheld, while qualifying any broader suggestion that nothing about the model was public [e1, e2].
Limitations
- The supplied excerpts do not directly establish the disclosure status of every listed category, particularly training energy and processor count.
- Withheld compute details would not by themselves establish withheld electricity consumption; a hardware description would not establish processor count.
- The March 15 sources do not establish their precise publication order relative to Marcus's post.
Evidence (2 items)
Supports the claim: The Verge,
And the reason we don’t disclose the training data is pretty much the same reason we don’t disclose the number of parameters.
Why it matters here: Sutskever directly confirms withholding size and training data. The article also documents the absent energy and hardware information and reproduces OpenAI's statement withholding architectural and training details.
- Publisher
- The Verge
- Kind of source
- News report
- Describes events of
- Public from
Context only: OpenAI,
GPT-4 is a Transformer- based model pre-trained to predict the next token in a document.
Why it matters here: The report disclosed the general architecture and broad data-source categories, while section 2 expressly withheld further architectural, hardware, compute, and dataset details. This qualifies any literal reading that nothing whatsoever was known.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, GPT-4’s successes, and GPT-4’s failures
We don’t know how big it is; we don’t know what the architecture is, we don’t know how much energy was used; we don’t how many processors were used; we don’t know what it was trained on etc.
- Where
- paragraph 8
- Published
- 2023-03-15
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-gpt-4s-successes-and-gpt-4s-failures-17c3ce-a2: Current. Outcome: Supported. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.
About assessment marcus-gpt-4s-successes-and-gpt-4s-failures-17c3ce-a2.