“And all this from a company thought (or claimed) that GPT-5 was going to be tantamount to AGI (spoiler alert: it wasn’t)”
Our reading, the sentence that was checked:
GPT-5 was not tantamount to AGI, even though OpenAI thought or claimed it would be.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Peak bubble, paragraph 7.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-peak-bubble-60174f-a1
We recorded 8 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2025-09-11. 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
- GPT-5 was going to be tantamount to AGI
- a company thought (or claimed)
- spoiler alert: it wasn’t
Our interpretation
- The claim contains two assertions: OpenAI expected or represented GPT-5 as effectively AGI, and the released product did not meet that description.
- The ordinary reading of AGI is broadly human-level general cognitive capability, rather than excellence on selected benchmarks. Marcus supplies no formal definition.
- The capability assessment concerns the August 2025 release through the post date. Earlier statements are relevant to the separate assertion about OpenAI's expectations.
- General predictions that OpenAI would eventually develop AGI do not establish a prediction specifically about GPT-5.
Counts as supported if: Evidence establishes both a GPT-5-specific expectation or representation by OpenAI and substantial shortcomings inconsistent with the ordinary reading of AGI at release.
Counts as contradicted if: Reliable evidence establishes that released GPT-5 met the stated ordinary reading of AGI, or directly refutes the particular prior expectation or representation attributed to OpenAI. Failure to find an earlier statement is not itself a contradiction.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Unresolved
The record was checked, and the evidence available by the cutoff does not settle it.
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
The non-AGI part has substantial support. Altman acknowledged that GPT-5 had not reached AGI, and an independent study documented important capability gaps. The separate assertion that OpenAI previously thought or claimed GPT-5 would be effectively AGI remains unverified. Supporting one part does not establish the whole statement, and the missing attribution is not evidence that Marcus was wrong.
Limitations
- The independent evaluation covers spatial intelligence, not every cognitive capability.
- The executive acknowledgment is reported by a reputable news outlet.
- Internal expectations may not be publicly documented.
Evidence (2 items)
Supports the claim: The Guardian,
However, Altman admitted GPT-5 had not reached that goal yet.
Why it matters here: Supports the non-AGI part: the surrounding passage identifies AGI as the goal and reports missing continuous learning. This is news reporting of Altman's remarks, not a directly retrieved company transcript, and does not establish his earlier expectation.
- Publisher
- The Guardian
- Kind of source
- News report
- Describes events of
- Public from
Supports the claim: arXiv; Zhongang Cai and coauthors,
Nevertheless, they continue to exhibit notable limitations in spatial understanding and reasoning, which are fundamental capabilities to achieving artificial general intelligence.
Why it matters here: The original study reports GPT-5 below human performance across a broad range of spatial tasks, while also recognizing improved capabilities. This supports substantial remaining limitations, but is neither a comprehensive AGI test nor evidence about OpenAI's prior beliefs.
- Publisher
- arXiv; Zhongang Cai and coauthors
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
What would settle it
A dated, attributable OpenAI statement or internal record specifically connecting GPT-5 to the alleged AGI expectation, alongside the release-period capability evidence.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Peak bubble
And all this from a company thought (or claimed) that GPT-5 was going to be tantamount to AGI (spoiler alert: it wasn’t)
- Where
- paragraph 7
- Published
- 2025-09-11
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-peak-bubble-60174f-a1: Current. Outcome: Unresolved. 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.