“The fact that GPT 4.5 received very mixed reviews — despite years of labor and billions investment — and could in no reasonable way be considered a major (rather than minor, incremental) step forwards towards AGI.”
Our reading, the sentence that was checked:
GPT-4.5 received very mixed reviews and is at most a minor, incremental step toward AGI, despite years of work and billions invested.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Ezra Klein’s new take on AGI – and why I think it’s probably wrong, paragraph 40.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-ezra-kleins-new-take-on-agi-and-why-9478c3-a1
We recorded 29 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 2025-03-05. 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)
Conditions
- despite years of labor and billions invested (qualifier, stated by the speaker)
From the speaker
- received very mixed reviews
- despite years of labor and billions investment
- could in no reasonable way be considered a major
- minor, incremental
- step forwards towards AGI
Our interpretation
- Mixed reviews means contemporaneous assessments included both praise and substantial criticism.
- The ordinary capability reading is that GPT-4.5 did not deliver a broad step change toward the post's definition of AGI; it does not require every individual benchmark gain to be small.
- The target period is the release and initial reception, February 27 through March 5, 2025.
- Years of labor and billions invested are additional factual assertions, not conditional antecedents.
- The spending assertion concerns cumulative GPT-4.5 development costs through release, not all OpenAI expenditure or the cost of a single final training run.
- USD is inferred from the surrounding discussion. The source supplies no identifiable accounting basis or development start date.
Counts as supported if: Contemporaneous reviews are materially divided, broad comparative evaluations support only incremental progress toward general cognitive competence, and model-specific records substantiate multiple years of development and billions in spending.
Counts as contradicted if: Reliable evidence refutes the mixed reception or development-resource assertions, or demonstrates a broad capability advance reasonably qualifying as a major step toward the stated AGI definition.
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 initial reception included praise and substantial criticism, and METR's tests placed GPT-4.5 between existing comparison models. OpenAI also reported substantial gains in particular domains. These findings do not settle the categorical judgment about progress toward AGI. Model-specific development spending and its accounting basis remain unverified, so the combined claim cannot receive a supported or contradicted outcome.
Limitations
- Company benchmarks and limited pre-release evaluations do not exhaust a model's capabilities.
- The available cost analysis covers a different period and broader spending scope.
- No agreed threshold separates major from incremental progress toward AGI.
Evidence (4 items)
Context only: Ars Technica,
these incremental gains
Why it matters here: The article offers a strongly critical assessment while also reporting Karpathy's favorable reaction. This supports the mixed-reception component; Marcus's own comments in the article are not used.
- Publisher
- Ars Technica
- Kind of source
- News report
- Describes events of
- Public from
Context only: METR,
between GPT-4o and o1.
Why it matters here: Pre-release autonomy and research-engineering tests did not place GPT-4.5 beyond the existing comparison models. METR cautions that its short evaluation of an earlier checkpoint may underestimate capabilities.
- Where
- Opening paragraph
- Publisher
- METR
- Kind of source
- Independent measurement
- Describes events of
- Public from
Context only: OpenAI,
meaningful improvements
Why it matters here: Reported GPQA scores rose from 53.6% to 71.4%, and AIME scores from 9.3% to 36.7%, compared with GPT-4o. These substantial domain gains challenge an unqualified dismissal, but do not establish a major advance toward AGI.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Context only: Epoch AI,
compute in 2024
Why it matters here: The analysis separates company-wide expenditure from final training runs and research experiments. Its aggregate estimate cannot substantiate billions spent specifically developing GPT-4.5.
- Publisher
- Epoch AI
- Kind of source
- Independent measurement
- Measures
- OpenAI, Company-wide cloud compute expenditure, USD billions, estimate, 2024-01-01 to 2024-12-31
- Describes events of
- Public from
What would settle it
Model-specific development and spending records, a broader contemporaneous review sample, and independent evaluations connecting the capability changes to an explicit, defensible AGI-progress criterion.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Ezra Klein’s new take on AGI – and why I think it’s probably wrong
The fact that GPT 4.5 received very mixed reviews — despite years of labor and billions investment — and could in no reasonable way be considered a major (rather than minor, incremental) step forwards towards AGI.
- Where
- paragraph 40
- Published
- 2025-03-05
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-ezra-kleins-new-take-on-agi-and-why-9478c3-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.
About assessment marcus-ezra-kleins-new-take-on-agi-and-why-9478c3-a1.