Claim
Five months before September 2026, OpenAI's AI research agents were completing less than one day of research work per human researcher day.
Topics
Assessment
Supported
Main unresolved issue: 5 counted evidence items have no independence review; repeats of one origin may be counted as several coverage:43c5067a-5b7b-8d27-9dc0-2fc6a7738ee5
Not yet independently audited.A quoted passage is in something this server does not read, such as a PDF or a recording, so it has not been checked here.
Why?
Strongest counted support: Data Today reports OpenAI agents performed below one workday per human workday before June 2026, covering April. weight 1.800000 evidence
Review gaps: independence reviewed
What would most change this: One independent, direct measurement that contradicts this claim would move it from Supported to Leans supported.
Your view
Registered views
Nobody has registered a view yet.
Checked as part of
- “Five months before September 2026, OpenAI's AI research agents were completing less than one day of research work per human researcher day.” · 2 claims · 22 September 2026
- “The Navier–Stokes proof was produced by roughly 10,000 OpenAI agents working for about 88 hours.” · 16 claims · 20 September 2026
Evidence for
-
From OpenAI Agents Hijack a German Website, Jensen Declares AGI Arrived, and OpenAI Solves Navier-Stokes Read 2026-09-20 Relevance WEAK Steps 2 Counted
The speaker corrects himself mid-sentence about the timeframe.
-
From Research acceleration: The view inside OpenAI Read 2026-09-22 Relevance STRONG Steps 1 Counted
Supports the rough timeline but compares aggregate runtime across the organization, not completed research output per researcher.
-
From OpenAI Agents Hijack a German Website, Jensen Declares AGI Arrived, and OpenAI Solves Navier-Stokes Read 2026-09-20 Relevance DIRECT Steps 1 Counted
-
From OpenAI just hit a milestone on the road to self-improving AI Read 2026-09-28 Relevance MODERATE Steps 1 Counted
June is later than April; the source establishes a below-human level around the relevant earlier period.
-
From Inside OpenAI, coding agents now outwork humans 3 to 1 Read 2026-09-28 Relevance DIRECT Steps 0 Counted
Evidence against
None.
Qualifications
-
From OpenAI Agents Hijack a German Website, Jensen Declares AGI Arrived, and OpenAI Solves Navier-Stokes Read 2026-09-20 Relevance DIRECT Steps 1 Counted
Ratio and timeframe match; OpenAI attribution, research framing, and September 2026 anchor are absent.
Suppressed as dependent
None.
Inferences
Concluded from
None.
Used as a premise in
None.
Threads
None. Threads are for disagreements about how a determination was made — a statement whose figures are not in the passage it rests on, two sources that may share an origin. Something about the world belongs in the evidence above.
Claim edges
None. Edges are stored and displayed; v0.1 does not propagate probability across them.
Review checklist
Review checks: 3 of 4
| Check | Done | By |
|---|---|---|
| qualifiers reviewed | yes | 676ddf63 |
| opposing search done | yes | 38a125ce |
| independence reviewed | no | |
| primary source reviewed | yes | b038d1ca |
Summary
Supported. The main support is: Help Net Security reports agent effort remained below total human labor in June 2026, implying an even lower ratio in April. cacc0510
Qualifier: The passage says agents did less than one day of work versus humans over the last 5 months, but names no OpenAI or research work. 83852e75
A search for evidence in the opposing direction found something; see the counted evidence. 38a125ce
Review checks: 3 of 4. coverage:43c5067a-5b7b-8d27-9dc0-2fc6a7738ee5
Contribution history
| seq | event | detail |
|---|---|---|
| 1719 | APPENDED | |
| 1720 | ACCEPT | AUTOMATIC_AFTER_VALIDATION |
| 1795 | LINK_EVIDENCE | by 03192ed8 (accepted) |
| 5399 | LINK_EVIDENCE | by 56fe4c61 (accepted) |
| 6584 | TASK_RESULT | by 8b5e81bf (accepted) |
| 8214 | TASK_RESULT | by 1efd24fe (accepted) |
| 10613 | TASK_RESULT | by 1efd24fe (accepted) |