“Hallucinations simply haven’t gone away, even in the latest models.”
Our reading, the sentence that was checked:
Hallucinations have not gone away, even in the latest models.
Assessment: Supported 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 21.
- 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-21d48a-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)
From the speaker
- Hallucinations simply haven’t gone away
- even in the latest models
Our interpretation
- The plain reading is that contemporary frontier models still sometimes present invented or incorrect factual information as fact.
- This does not assert a particular hallucination rate, equal rates across models, or an absence of improvement.
- For the ongoing state, use early-2025 evaluations available through the post date, January 1–March 5, 2025. AI capability evaluations have no standard reporting quarter.
Counts as supported if: A documented evaluation of a contemporary frontier model reports a nonzero rate of factual hallucinations.
Counts as contradicted if: Credible evaluations establish that contemporary frontier models have eliminated factual hallucinations across relevant uses.
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
OpenAI’s February 2025 release evaluation reports a nonzero hallucination rate for GPT-4.5. That directly supports the ordinary reading that hallucinations persisted in contemporary frontier models. The lower rate than the comparison models also shows why persistence should not be read as an absence of progress.
Limitations
- PersonQA concerns factual questions about people; its rate is not an estimate for all uses.
- These are company-reported evaluations, not an independent replication.
Evidence (1 item)
Supports the claim: OpenAI,
aims to elicit hallucinations.
Why it matters here: Table 4 reports PersonQA hallucination rates of 0.19 for GPT-4.5, 0.30 for GPT-4o, and 0.20 for o1. The release evaluation directly establishes persistence, while also showing improvement over the comparison models.
- 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, Ezra Klein’s new take on AGI – and why I think it’s probably wrong
Hallucinations simply haven’t gone away, even in the latest models.
- Where
- paragraph 21
- 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-21d48a-a1: 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-ezra-kleins-new-take-on-agi-and-why-21d48a-a1.