Claim

Training GPT-6 Astra on about 100,000 chips amounted to roughly a billion-dollar training run over about two months.

Type QUANTITATIVE · truth-evaluable: yes · status at snapshot: ACTIVE · created seq 1699

Referenced 215 times: checked 4, looked up by assistants 0, viewed 134, shared 77. A count of attention, not evidence.

Topics

AI · 03192ed8 Business · 03192ed8

Assessment

Contradicted

Score: 0.1520 under ledger-default@0.3.0 at snapshot 27745

2 of 4 review checks done · 2 counted

3 independent evidence lineages · review checks 2 of 4 · stability MEDIUM

Main unresolved issue: 7 counted evidence items have no independence review; repeats of one origin may be counted as several coverage:4c5f8598-33e4-8c60-ba5d-d4149e643708

Not yet independently audited.Evidence points both ways.A quoted passage from boaoai.cn and wccftech.com was not found on the page when Galedra fetched it.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: Greg Brockman said GPT-6 Astra was OpenAI's first training run using more than 100,000 GPUs. weight 0.540000 evidence

Strongest counted contradiction: The report supports a training run exceeding 100,000 GPUs and an estimated cost range reaching $1 billion, but dates the run at 117 days (roughly four months), not about two months. weight 1.800000 evidence

Review gaps: independence reviewed, primary source reviewed

What would most change this: One independent, direct measurement that supports this claim would move it from Contradicted to Unresolved.

Show calculation

Your view

Registered views

Nobody has registered a view yet.

Registered views are personal belief, not evidence: they never change Galedra's assessment. Affiliations are self-declared and shown only for groups of 5 or more.

Checked as part of

Statements this claim was broken out of. The claim is scored on its own; a statement's reading only counts its claims.

“Training GPT-6 Astra on about 100,000 chips amounted to roughly a billion-dollar training run over about two months.” · 1 claim · 28 September 2026
“Training GPT-6 Astra on about 100,000 chips amounted to roughly a billion-dollar training run over about two months.” · 1 claim · 28 September 2026
“Training GPT-6 Astra on about 100,000 chips amounted to roughly a billion-dollar training run over about two months.” · 1 claim · 28 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

Mostaque estimates the 100,000-chip run at roughly a billion dollars over two months.

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

An off-the-cuff estimate, not a disclosed figure.
Greg Brockman said GPT-6 Astra was OpenAI's first training run using more than 100,000 GPUs.

From Jensen Huang says 100,000 Nvidia GPUs were used to train OpenAI's latest model, GPT-6 Astra Read 2026-09-28 Relevance MODERATE Steps 0 Counted

Supports the approximately 100,000-chip scale, but does not establish the claimed billion-dollar cost or two-month duration.

Evidence against

The report supports a training run exceeding 100,000 GPUs and an estimated cost range reaching $1 billion, but dates the run at 117 days (roughly four months), not about two months.

From OpenAI GPT-6 Astra Enters the 'AGI Era' Read 2026-09-28 Relevance DIRECT Steps 0 Counted

The cost is an estimate based on assumptions rather than an official disclosure; the duration directly conflicts with the claim's two-month figure.
Outside reporting places the assumed 100,000-GPU Astra run at roughly 90–120 days, not two months, while making clear that the $500 million–$1 billion figure is a rate-based estimate rather than a disclosed cost.

From OpenAI’s ~$1 Billion GPT-6 Astra Heralds The Age Of AGI Read 2026-09-28 Relevance MODERATE Steps 1 Counted

Directly addresses the duration and cost components while preserving the source's assumptions.

Qualifications

The passage cites 100,000 chips, a roughly billion-dollar run over about 2 months, but names no model.

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

Chips, cost, and duration match; 'GPT-6 Astra' is unnamed in the passage.
The excerpt says, with uncertainty, that 100,000 chips were used and characterizes it as roughly a billion-dollar training run over roughly two months. It directly supports the figures but preserves the speaker's uncertainty about the chip count.

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
CNET reports OpenAI pretrained Astra on more than 100,000 GPUs in its largest-scale training run.

From OpenAI’s Astra Is Here: What to Know About GPT-6 Read 2026-09-28 Relevance STRONG Steps 0 Counted

Directly supports the 100,000-GPU component, but the article does not establish a billion-dollar cost or two-month duration; those remain estimates.

Suppressed as dependent

None.

Inferences

Concluded from

None.

Used as a premise in

None.

An inference is interpretation, not evidence: it never changes the assessment of its conclusion. Each premise is a claim with its own evidence.

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: 2 of 4

2 checks by this claim's own author, alongside the independent review above. 1 QUALIFIER CHECK · 1 EVIDENCE VERIFICATION. A check by the person who recorded a claim does not raise its review coverage.

Check	Done	By
qualifiers reviewed	yes	630d14af
opposing search done	yes	c775d8ab
independence reviewed	no
primary source reviewed	no

Summary

Contradicted. The main support is: Greg Brockman said GPT-6 Astra was OpenAI's first training run using more than 100,000 GPUs. 790e7f79

Against that, contradiction: Outside reporting places the assumed 100,000-GPU Astra run at roughly 90–120 days, not two months, while making clear that the $500 million–$1 billion figure is a rate-based estimate rather than a disclosed cost. 7e338c5e

Qualifier: The excerpt says, with uncertainty, that 100,000 chips were used and characterizes it as roughly a billion-dollar training run over roughly two months. It directly supports the figures but preserves the speaker's uncertainty about the chip count. 2e755ea3

Qualifier: The passage cites 100,000 chips, a roughly billion-dollar run over about 2 months, but names no model. 3dbb91a4

Qualifier: CNET reports OpenAI pretrained Astra on more than 100,000 GPUs in its largest-scale training run. 110f1417

A search for evidence in the opposing direction found something; see the counted evidence. c775d8ab

Generated by stub-v0.1; every sentence cites graph ids.

Contribution history

seq	event	detail
1699	APPENDED
1700	ACCEPT	AUTOMATIC_AFTER_VALIDATION
1785	LINK_EVIDENCE	by 03192ed8 (accepted)
6526	TASK_RESULT	by 8b5e81bf (accepted)
8120	TASK_RESULT	by 1efd24fe (accepted)
9607	LINK_EVIDENCE	by 1efd24fe (accepted)
9655	LINK_EVIDENCE	by 1efd24fe (accepted)
10498	TASK_RESULT	by 1efd24fe (accepted)
12302	LINK_EVIDENCE	by 1efd24fe (accepted)