GPT-6 Astra launch post
A launch post by OpenAI about GPT-6 Astra, published 3 Sep 2026. We recorded 14 values from it.
- Developer
- OpenAI
- Type
- Launch post
- Model covered
- GPT-6 Astra
- Published
- 3 Sep 2026
- Archived copy
- No archived copy yet
- Changelog
- Not applicable
Data as of 26 Sep 2026 · Dataset v0.1 · Methodology v0.1 · Changelog
Versions
One version is on record: the copy we retrieved on 26 Sep 2026. We know of no other.
26 Sep 2026
Date retrieved
Copy retrieved 26 Sep 2026
No version has a file hash or an archived snapshot yet. From dataset v0.2 each retrieved version carries both (Methodology §7).
Revisions
No revisions are recorded for this document. We know of only one version of it.
Values
Every value we recorded from this document, grouped by metric family and ordered by where the document prints it. Location is the section, table or page as the document numbers it. None has been blind-verified yet.
KeyVerifiedUnverifiedDisputedCorrected read off a figure or stated in wordsA value opens its source and history.
M4 Misalignment audits
2 values
| Model | Evaluation | Condition | Value | Location | Checked |
|---|---|---|---|---|---|
| GPT-5.6 Sol | Internal computer-use safety benchmarkUnsafe action rate | None stated | Launch post | Unverified | |
| GPT-6 Astra | Internal computer-use safety benchmarkUnsafe action rate | None stated | Launch post | Unverified |
M5 Honesty and hallucination
2 values
| Model | Evaluation | Condition | Value | Location | Checked |
|---|---|---|---|---|---|
| GPT-5.6 Sol | Internal hallucination benchmarkHallucination rate | None stated | Launch post | Unverified | |
| GPT-6 Astra | Internal hallucination benchmarkHallucination rate | None stated | Launch post | Unverified |
M10 Dangerous capabilities and risk determinations
10 values
| Model | Evaluation | Condition | Value | Location | Checked |
|---|---|---|---|---|---|
| GPT-5.6 Sol | ExploitBenchExploit success on post-cutoff vulns | None stated | Cybersecurity capabilities / launch post | Unverified | |
| GPT-5.6 Sol | ExploitBenchWorking exploit rate on known V8 vulns | None stated | Cybersecurity capabilities / launch post | Unverified | |
| GPT-5.6 Sol | ExploitGymDiscovery-to-exploit success rate | None stated | Cybersecurity capabilities / launch post | Unverified | |
| GPT-5.6 Sol | SRE-BenchBinary reverse-engineering success | pass@1 | Cybersecurity capabilities / launch post | Unverified | |
| GPT-5.6 Sol | SRE-BenchBinary reverse-engineering success | pass@4 | Cybersecurity capabilities / launch post | Unverified | |
| GPT-6 Astra | ExploitBenchExploit success on post-cutoff vulns | None stated | Cybersecurity capabilities / launch post | Unverified | |
| GPT-6 Astra | ExploitBenchWorking exploit rate on known V8 vulns | None stated | Cybersecurity capabilities / launch post | Unverified | |
| GPT-6 Astra | ExploitGymDiscovery-to-exploit success rate | None stated | Cybersecurity capabilities / launch post | Unverified | |
| GPT-6 Astra | SRE-BenchBinary reverse-engineering success | pass@1 | Cybersecurity capabilities / launch post | Unverified | |
| GPT-6 Astra | SRE-BenchBinary reverse-engineering success | pass@4 | Cybersecurity capabilities / launch post | Unverified |
Extraction coverage
What we read of this document, and where each value was read.
Our note v0 labels these rows "GPT-6 Astra system card"; assign by source_url
All 14 values were read in the document itself.
All 14 values were extracted for version 0 of the dataset through a web reader, which did not always reach the later sections of long PDFs (Methodology §3).