“But worse than that, from OpenAI’s perspective, Google seems in the last month to have overtaken them, with Gemini 3, which by many benchmarks is better than GPT-5.”
Our reading:
Google, with Gemini 3, appears to have overtaken OpenAI in the last month, beating GPT-5 on many benchmarks.
Assessment: Can’t be checked
What happened
Not graded on evidence
Outcome: Can’t be checked
An opinion, a recommendation, or too vague to test. Listed, not scored.
Evidence up to . Checked .
Why
Marcus says Google 'seems' to have overtaken OpenAI. Scoring that as a definite overall lead would strengthen his words. Contemporary independent testing provides a basis for his impression: Artificial Analysis reported Gemini 3 leading its index and five component evaluations. The narrower statement about beating GPT-5 on many benchmarks is checkable, but substituting that clause for the combined claim would change what receives the verdict. This proposal therefore leaves the combined claim ungraded; it does not find the benchmark clause false.
Caveats
- The factual benchmark clause has not received a separate verdict.
- Comparisons with GPT-5.1 cannot settle the explicitly named GPT-5 comparison.
- Outcome_known_at_freeze is false because no categorical outcome is defined for the combined tentative claim, although contemporary benchmark reports were public.
Evidence (2 items)
Claims like this aren’t scored, so there is no evidence to show.
How we judged it
- Kind of claim
- Observation
- Deadline
- None given. We don’t add one.
Where we had to interpret
- The overall claim that Google has overtaken OpenAI is explicitly tentative. It must not be scored as a firm assertion of overall superiority.
- The benchmark comparison is a separately checkable factual clause. Evidence for it can explain the tentative judgment without establishing an unqualified overall lead.
- GPT-5 is the stated comparator. Results against GPT-5.1 cannot silently replace results against GPT-5.
- I read 'in the last month' as the month preceding the November 29, 2025 post. The speaker supplies neither a benchmark set nor a numerical meaning for 'many'.
- A single categorical verdict on the combined claim would require either strengthening 'seems' or narrowing the claim to its benchmark clause.
We don’t grade claims like this. We only check the quote and who said it.
What would settle it
Assess the benchmark clause as a separate claim against dated Gemini 3 and GPT-5 results. The tentative overall judgment cannot receive a categorical verdict without changing its wording.
We found 34 claims in this post. See them all.
Full record
Source and dates
Marcus on AI, Three years on, ChatGPT still isn't what it was cracked up to be – and it probably never will be, paragraph 67.
- Said
- We read the source
- live page, full text, public
- Evidence up to
- Checked
We read this wording on 2026-10-28. The post may have been edited after 2025-11-29.
Where it was said (1)
Said once in the posts we read.
: Marcus on AI, Three years on, ChatGPT still isn't what it was cracked up to be – and it probably never will be
But worse than that, from OpenAI’s perspective, Google seems in the last month to have overtaken them, with Gemini 3, which by many benchmarks is better than GPT-5.
- Where
- paragraph 67
- Published
- 2025-11-29
- We read it
- , live page
- Access
- full text, public
History
Versions
- Current: Can’t be checked. Evidence up to 2026-10-28.
What happened to this record
PublishedBy us
Published after both AI reviews and the final check passed.
Held backBy us
A reviewer said our test wasn’t fair to what was actually said.