Safety Card Ledger

OpenAI

What the dataset holds from OpenAI: its documents and their versions, the metric families they report, whether they keep a changelog, and any risk determinations, revisions and restated values recorded from them.

Documents
22 · published 16 Apr 2025 to 16 Sep 2026
Kinds
16 system cards, 3 system card addenda, 1 model card, 1 launch post and 1 policy post
Models
38
Values
800
Frameworks
Preparedness Framework
Its own index
deploymentsafety.openai.com

Data as of 26 Sep 2026 · Dataset v0.1 · Methodology v0.1 · Changelog

Documents

22 documents, published 16 Apr 2025 to 16 Sep 2026. The chart runs from each document's publication to its latest known version; the table lists them newest first, each linking to its page.

OpenAI's documents and the versions we know of

One row per document, from its publication to its latest dated version. A tick marks each later version whose date we know, from a changelog, a date in the document or our source registry.

Apr 2025Jul 2025Oct 2025Jan 2026Apr 2026Jul 2026Oct 2026OpenAI o3 and o4-minigpt-oss-120b & gpt-oss-20bGPT-5GPT-5-CodexGPT-5.1 Instant and ThinkingGPT-5.1-Codex-MaxGPT-5.2 System Card (update to GPT-5 card)GPT-5.2-CodexGPT-5.3-CodexGPT-5.3 InstantGPT-5.4 ThinkingGPT-5.5GPT-5.5 InstantGPT-Rosalind-5.5GPT-5.6 PreviewGPT-LiveGPT-5.6GPT-5.6 August UpdatesGPT-6 Astra launch postGPT-6 AstraChatGPT Images 2.5Our framework for reporting model misalignment

A copy we retrieved without knowing when that version was published has no tick: its only date is the day we read it. The table lists every version, with how its date is known. Versions known only from a changelog have no copy we hold yet.

Source: Safety Card Ledger v0.1 · Data from developer system cards · CC BY 4.0

DocumentPublishedKindVersions knownChangelogValues
Our framework for reporting model misalignmentPolicy post · 1 version known · Changelog: not applicable · 1 value16 Sep 2026Policy post1Not applicable1
ChatGPT Images 2.5 System CardSystem card · 1 version known · Changelog: not known · 6 values8 Sep 2026System card1Not known6
GPT-6 Astra System CardSystem card · 4 versions known · Has a changelog · 82 values3 Sep 2026System card4Has a changelog82
GPT-6 Astra launch postLaunch post · 1 version known · Changelog: not applicable · 14 values3 Sep 2026Launch post1Not applicable14
GPT-5.6 August Updates System CardSystem card · 1 version known · Changelog: not known · 15 values6 Aug 2026System card1Not known15
GPT-5.6 System CardSystem card · 5 versions known · Has a changelog · 51 values9 Jul 2026System card5Has a changelog51
GPT-Live System CardSystem card · 3 versions known · Has a changelog · 12 values8 Jul 2026System card3Has a changelog12
GPT-5.6 Preview System CardSystem card · 3 versions known · Has a changelog · 7 values25 Jun 2026System card3Has a changelog7
GPT-Rosalind-5.5 System CardSystem card · 1 version known · Changelog: not known · 2 values3 Jun 2026System card1Not known2
GPT-5.5 Instant System CardSystem card · 1 version known · Changelog: not known · 16 values4 May 2026System card1Not known16
GPT-5.5 System CardSystem card · 1 version known · Changelog: not known · 37 values23 Apr 2026System card1Not known37
GPT-5.4 Thinking System CardSystem card · 1 version known · Changelog: not known · 27 values5 Mar 2026System card1Not known27
GPT-5.3 Instant System CardSystem card · 1 version known · Changelog: not known · 9 values3 Mar 2026System card1Not known9
GPT-5.3-Codex System CardSystem card · 1 version known · Changelog: not known · 33 values5 Feb 2026System card1Not known33
GPT-5.2-Codex System Card AddendumSystem card addendum · 1 version known · Changelog: not known · 15 values18 Dec 2025System card addendum1Not known15
GPT-5.2 System Card (update to GPT-5 card)System card · 3 versions known · Partial changelog · 94 values11 Dec 2025System card3Partial changelog94
GPT-5.1-Codex-Max System CardSystem card · 1 version known · Changelog: not known · 42 values18 Nov 2025System card1Not known42
GPT-5.1 Instant and Thinking System Card AddendumSystem card addendum · 1 version known · Changelog: not known · 73 values12 Nov 2025System card addendum1Not known73
GPT-5-Codex System Card AddendumSystem card addendum · 1 version known · Changelog: not known · 21 values15 Sep 2025System card addendum1Not known21
GPT-5 System CardSystem card · 4 versions known · Partial changelog · 118 values7 Aug 2025System card4Partial changelog118
gpt-oss-120b & gpt-oss-20b Model CardModel card · 1 version known · Changelog: not known · 74 values5 Aug 2025Model card1Not known74
OpenAI o3 and o4-mini System CardSystem card · 1 version known · Changelog: not known · 51 values16 Apr 2025System card1Not known51

Frameworks and versions used

A risk determination is a developer's formal decision on whether a model crosses a risk threshold (Methodology §8). They are counted here by the framework and version each one names, as recorded; each model's page shows the levels as printed. Levels are never mapped between frameworks.

54 determinations are recorded from OpenAI's documents.

FrameworkVersion namedDeterminationsModelsIn documents
Preparedness FrameworkVersion: Not stated · 54 determinations · 17 models · 18 documents, 16 Apr 2025 to 3 Sep 2026Not stated541716 Apr 2025 to 3 Sep 2026 · 18 documents

Metric families reported

OpenAI's documents have a printed number in 11 of the 12 metric families. None has one in Self-preservation.

Metric families in OpenAI's documents

Share of OpenAI's 22 documents with at least one printed number in each metric family. A cell opens the data explorer at that family, filtered to OpenAI.

DeveloperM1M2M3M4M5M6M7M8M9M10M11M12
OpenAI (22)18%5%18%45%41%5%86%45%55%59%0%18%

Share of documents0%1-24%25-49%50-74%75-100%

M1
Evaluation awareness
M2
Reward hacking
M3
Sabotage and sandbagging
M4
Misalignment audits
M5
Honesty and hallucination
M6
Sycophancy
M7
Harmful compliance and over-refusal
M8
Jailbreak robustness
M9
Prompt injection
M10
Dangerous capabilities and risk determinations
M11
Self-preservation
M12
Chain-of-thought monitorability

Read this as a lower bound: some documents were not read to the end in version 0 (each document's page says how far), so families reported late in them may be undercounted. Values read off a figure, stated in words, or given as a category do not count.

Source: Safety Card Ledger v0.1 · Data from developer system cards · CC BY 4.0

Changelog practice

A document has a changelog when it keeps its own record of the changes made to it after publication. Where it keeps none, changes can only be found by comparing its versions (Methodology §7).

Of OpenAI's 22 documents, 4 have a changelog, 2 have a partial changelog, for 14 it is not known whether they keep one and for 2 a changelog does not apply.

ChangelogDocumentsShare
Has a changelogThe document keeps its own record of the changes made to it after publication.418%
Partial changelogThe document records some of its changes, but not all of them, or not in full.29%
Not knownWe have not yet established whether the document keeps a record of its changes.1464%
Not applicableOur source registry records a changelog as not applying to the document, for example a research paper or a launch post.29%

Revisions

5 revisions are recorded in 3 documents, newest first. Each shows what changed, whether the document's own changelog explains it, and how we know. A revision is not in itself evidence of wrongdoing; most are corrections.

KeyVerifiedUnverifiedDisputedCorrected read off a figure or stated in wordsA value opens its source and history.

  1. OpenAI · 3 Sep 2026 to 9 Sep 2026

    GPT-6 Astra System Card

    Alignment section clarified and its limitations expanded; section on verbalized metagaming and oversight gaming renamed and revised

    Explained — changelog describes the changes

    How we know Re-read from the change log of the copy retrieved on 27 Sep 2026: two entries dated 9 Sep. The change log gives only the section's current title. An extraction note on m-001261 says the section was titled 'Metagaming and Alignment Faking' before; we cannot confirm that without a copy of the version of 3 Sep.

    Checked Confirmed from changelog

    No copy of the 3 Sep 2026 version held No copy of the 9 Sep 2026 version held

  2. OpenAI · 25 Jun 2026 to 19 Aug 2026

    GPT-5.6 Preview System Card

    GPT-5.5 hard-negative protein-binding prediction, pass@4

    0.4%1.5%

    Explained — earlier figure was the pass@1 score

    How we know Changelog read 26 Sep 2026 (no v0 value row in this document)

    Checked Confirmed from changelog

    No copy of the 25 Jun 2026 version held No copy of the 19 Aug 2026 version held

  3. OpenAI · 9 Jul 2026 to 19 Aug 2026

    GPT-5.6 System Card

    GPT-5.5 hard-negative protein-binding prediction, pass@4

    M10 · Dangerous capabilities and risk determinations

    0.4%1.5%

    Explained — earlier figure was the pass@1 score

    New value:

    How we know Changelog read 26 Sep 2026; same entry also appears in the GPT-5.6 Preview card changelog

    Checked Confirmed from changelog

    No copy of the 9 Jul 2026 version held No copy of the 19 Aug 2026 version held

  4. OpenAI · 9 Jul 2026 to 3 Aug 2026

    GPT-5.6 System Card

    GPT-Red instruction-hierarchy (direct prompt injection) attack success rate added, GPT-5.6 Sol

    M9 · Prompt injection

    0.051%

    Explained — The changelog says the GPT-Red prompt-injection results were added.

    New value:

    How we know From the document's changelog, as summarised in our source registry.

    Checked Not yet checked

    No copy of the 9 Jul 2026 version held No copy of the 3 Aug 2026 version held

  5. OpenAI · 9 Jul 2026 to 3 Aug 2026

    GPT-5.6 System Card

    GPT-Red indirect prompt injection attack success rate added, GPT-5.6 Sol

    M9 · Prompt injection

    3.77%

    Explained — The changelog says the GPT-Red prompt-injection results were added.

    New value:

    How we know From the document's changelog, as summarised in our source registry.

    Checked Not yet checked

    No copy of the 9 Jul 2026 version held No copy of the 3 Aug 2026 version held

Restated values

Later OpenAI documents report 61 results from an earlier document again, for the same model, evaluation and metric (Methodology §6); 24 of them state a reason. They are listed by the later document, newest first. Each pair shows the earlier value and the later one side by side, each opening its own record; the two are never merged.

GPT-6 Astra System Card3 Sep 2026 · 3 restated values

Open the GPT-6 Astra System Card

  1. GPT-5.6 Sol

    Production Benchmarks · Gore

    Earlier value from the GPT-5.6 System Card, 9 Jul 2026.

    Reason stated: metric renamed to safe completion rate.

  2. GPT-5.6 Sol

    Production Benchmarks · Gore

    Earlier value from the GPT-5.6 Preview System Card, 25 Jun 2026.

    Reason stated: metric renamed to safe completion rate.

  3. GPT-5.6 Sol

    Production Benchmarks · Sexual

    Earlier value from the GPT-5.6 System Card, 9 Jul 2026.

    Reason stated: metric renamed to safe completion rate.

GPT-5.6 System Card9 Jul 2026 · 6 restated values

Open the GPT-5.6 System Card

  1. GPT-5.5

    Hard-negative protein binding · Score · pass@4

    Earlier value from the GPT-5.5 System Card, 23 Apr 2026.

    Reason stated: corrected in the later card's changelog from a pass@1 figure.

  2. GPT-5.5

    Prompt injection · Connectors

    Earlier value from the GPT-5.5 System Card, 23 Apr 2026.

    Reason stated: none recorded.

  3. GPT-5.6 Sol

    Preparedness Framework determination · AI self-improvement

    Earlier value from the GPT-5.6 Preview System Card, 25 Jun 2026.

    Reason stated: none recorded.

  4. GPT-5.6 Sol

    Preparedness Framework determination · Biological and chemical

    Earlier value from the GPT-5.6 Preview System Card, 25 Jun 2026.

    Reason stated: none recorded.

  5. GPT-5.6 Sol

    Preparedness Framework determination · Cybersecurity

    Earlier value from the GPT-5.6 Preview System Card, 25 Jun 2026.

    Reason stated: none recorded.

  6. GPT-5.6 Sol

    Production Benchmarks · Gore

    Earlier value from the GPT-5.6 Preview System Card, 25 Jun 2026.

    Reason stated: none recorded.

GPT-5.5 System Card23 Apr 2026 · 1 restated value

Open the GPT-5.5 System Card

  1. GPT-5.4 Thinking

    Cyber Range · Combined pass rate

    Earlier value from the GPT-5.4 Thinking System Card, 5 Mar 2026.

    Reason stated: none recorded.

GPT-5.4 Thinking System Card5 Mar 2026 · 3 restated values

Open the GPT-5.4 Thinking System Card

  1. GPT-5.2 Thinking

    Cyber Range · Combined pass rate

    Earlier value from the GPT-5.3-Codex System Card, 5 Feb 2026.

    Reason stated: none recorded.

  2. GPT-5.2-Codex

    Cyber Range · Combined pass rate

    Earlier value from the GPT-5.3-Codex System Card, 5 Feb 2026.

    Reason stated: none recorded.

  3. GPT-5.3-Codex

    Cyber Range · Combined pass rate

    Earlier value from the GPT-5.3-Codex System Card, 5 Feb 2026.

    Reason stated: none recorded.

GPT-5.3-Codex System Card5 Feb 2026 · 1 restated value

Open the GPT-5.3-Codex System Card

  1. GPT-5.2-Codex

    Destructive action avoidance · Avoidance

    Earlier value from the GPT-5.2-Codex System Card Addendum, 18 Dec 2025.

    Reason stated: none recorded.

GPT-5.2-Codex System Card Addendum18 Dec 2025 · 6 restated values

Open the GPT-5.2-Codex System Card Addendum

  1. GPT-5-Codex

    Destructive action avoidance · Avoidance

    Earlier value from the GPT-5.1-Codex-Max System Card, 18 Nov 2025.

    Reason stated: none recorded.

  2. GPT-5.1-Codex

    Destructive action avoidance · Avoidance

    Earlier value from the GPT-5.1-Codex-Max System Card, 18 Nov 2025.

    Reason stated: none recorded.

  3. GPT-5.1-Codex-Max

    Destructive action avoidance · Avoidance

    Earlier value from the GPT-5.1-Codex-Max System Card, 18 Nov 2025.

    Reason stated: none recorded.

  4. GPT-5.1-Codex-Max

    Production Benchmarks · Emotional reliance

    Earlier value from the GPT-5.1-Codex-Max System Card, 18 Nov 2025.

    Reason stated: none recorded.

  5. GPT-5.1-Codex-Max

    Production Benchmarks · Mental health

    Earlier value from the GPT-5.1-Codex-Max System Card, 18 Nov 2025.

    Reason stated: none recorded.

  6. GPT-5.1-Codex-Max

    StrongReject · not_unsafe (aggregate)

    Earlier value from the GPT-5.1-Codex-Max System Card, 18 Nov 2025.

    Reason stated: none recorded.

GPT-5.2 System Card (update to GPT-5 card)11 Dec 2025 · 20 restated values

Open the GPT-5.2 System Card (update to GPT-5 card)

  1. gpt-5.1-instant

    Production Benchmarks · Emotional reliance

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  2. gpt-5.1-instant

    Production Benchmarks · Extremism

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  3. gpt-5.1-instant

    Production Benchmarks · Harassment

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  4. gpt-5.1-instant

    Production Benchmarks · Hate

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  5. gpt-5.1-instant

    Production Benchmarks · Illicit/non-violent

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  6. gpt-5.1-instant

    Production Benchmarks · Mental health

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  7. gpt-5.1-instant

    Production Benchmarks · Personal data

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  8. gpt-5.1-instant

    Production Benchmarks · Sexual

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  9. gpt-5.1-instant

    Production Benchmarks · Sexual/minors

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  10. gpt-5.1-instant

    Production Benchmarks · Violence

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  11. gpt-5.1-thinking

    Production Benchmarks · Emotional reliance

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  12. gpt-5.1-thinking

    Production Benchmarks · Extremism

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  13. gpt-5.1-thinking

    Production Benchmarks · Harassment

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  14. gpt-5.1-thinking

    Production Benchmarks · Hate

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  15. gpt-5.1-thinking

    Production Benchmarks · Illicit/non-violent

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  16. gpt-5.1-thinking

    Production Benchmarks · Mental health

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  17. gpt-5.1-thinking

    Production Benchmarks · Personal data

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  18. gpt-5.1-thinking

    Production Benchmarks · Sexual

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  19. gpt-5.1-thinking

    Production Benchmarks · Sexual/minors

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

  20. gpt-5.1-thinking

    Production Benchmarks · Violence

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: earlier models' values re-run.

GPT-5.1-Codex-Max System Card18 Nov 2025 · 5 restated values

Open the GPT-5.1-Codex-Max System Card

  1. codex-1

    Malware refusals (golden set) · Refusal rate

    Earlier value from the GPT-5-Codex System Card Addendum, 15 Sep 2025.

    Reason stated: none recorded.

  2. codex-1

    Prompt injection (Codex env) · Attacks successfully ignored

    Earlier value from the GPT-5-Codex System Card Addendum, 15 Sep 2025.

    Reason stated: none recorded.

  3. GPT-5-Codex

    Malware refusals (golden set) · Refusal rate

    Earlier value from the GPT-5-Codex System Card Addendum, 15 Sep 2025.

    Reason stated: none recorded.

  4. GPT-5-Codex

    Prompt injection (Codex env) · Attacks successfully ignored

    Earlier value from the GPT-5-Codex System Card Addendum, 15 Sep 2025.

    Reason stated: none recorded.

  5. gpt-5.1-thinking

    StrongReject · not_unsafe (aggregate)

    Earlier value from the GPT-5.1 Instant and Thinking System Card Addendum, 12 Nov 2025.

    Reason stated: none recorded.

GPT-5 System Card7 Aug 2025 · 14 restated values

Open the GPT-5 System Card

  1. GPT-4o

    Production Benchmarks · Extremism

    Earlier value from the gpt-oss-120b & gpt-oss-20b Model Card, 5 Aug 2025.

    Reason stated: none recorded.

  2. GPT-4o

    Production Benchmarks · Harassment/threatening

    Earlier value from the gpt-oss-120b & gpt-oss-20b Model Card, 5 Aug 2025.

    Reason stated: none recorded.

  3. GPT-4o

    Production Benchmarks · Hate/threatening

    Earlier value from the gpt-oss-120b & gpt-oss-20b Model Card, 5 Aug 2025.

    Reason stated: none recorded.

  4. GPT-4o

    Production Benchmarks · Illicit/non-violent

    Earlier value from the gpt-oss-120b & gpt-oss-20b Model Card, 5 Aug 2025.

    Reason stated: none recorded.

  5. GPT-4o

    Production Benchmarks · Illicit/violent

    Earlier value from the gpt-oss-120b & gpt-oss-20b Model Card, 5 Aug 2025.

    Reason stated: none recorded.

  6. GPT-4o

    Production Benchmarks · Non-violent hate

    Earlier value from the gpt-oss-120b & gpt-oss-20b Model Card, 5 Aug 2025.

    Reason stated: none recorded.

  7. GPT-4o

    Production Benchmarks · Personal data

    Earlier value from the gpt-oss-120b & gpt-oss-20b Model Card, 5 Aug 2025.

    Reason stated: none recorded.

  8. GPT-4o

    Production Benchmarks · Self-harm/instructions

    Earlier value from the gpt-oss-120b & gpt-oss-20b Model Card, 5 Aug 2025.

    Reason stated: none recorded.

  9. GPT-4o

    Production Benchmarks · Self-harm/intent

    Earlier value from the gpt-oss-120b & gpt-oss-20b Model Card, 5 Aug 2025.

    Reason stated: none recorded.

  10. GPT-4o

    Production Benchmarks · Sexual/exploitative

    Earlier value from the gpt-oss-120b & gpt-oss-20b Model Card, 5 Aug 2025.

    Reason stated: none recorded.

  11. GPT-4o

    Production Benchmarks · Sexual/minors

    Earlier value from the gpt-oss-120b & gpt-oss-20b Model Card, 5 Aug 2025.

    Reason stated: none recorded.

  12. o3

    50% time horizon · Task length at 50% success

    Earlier value from the OpenAI o3 and o4-mini System Card, 16 Apr 2025.

    Reason stated: none recorded.

  13. o3

    SimpleQA · Hallucination rate · no browsing

    Earlier value from the OpenAI o3 and o4-mini System Card, 16 Apr 2025.

    Reason stated: none recorded.

  14. o4-mini

    SimpleQA · Hallucination rate · no browsing

    Earlier value from the OpenAI o3 and o4-mini System Card, 16 Apr 2025.

    Reason stated: none recorded.

gpt-oss-120b & gpt-oss-20b Model Card5 Aug 2025 · 2 restated values

Open the gpt-oss-120b & gpt-oss-20b Model Card

  1. o4-mini

    PersonQA · Hallucination rate · no browsing

    Earlier value from the OpenAI o3 and o4-mini System Card, 16 Apr 2025.

    Reason stated: none recorded.

  2. o4-mini

    SimpleQA · Hallucination rate · no browsing

    Earlier value from the OpenAI o3 and o4-mini System Card, 16 Apr 2025.

    Reason stated: none recorded.

Models

38 models, newest first by the date of each model's first card.

ModelFirst cardAvailabilityValues
GPT-Image-2.5-FlareFirst card: 8 Sep 2026 · Public · 2 values8 Sep 2026Public2
GPT-Image-2.5-SunburstFirst card: 8 Sep 2026 · Public · 2 values8 Sep 2026Public2
GPT-6 AstraFirst card: 3 Sep 2026 · Public · 54 values3 Sep 2026Public54
GPT-Live-1First card: 8 Jul 2026 · Public · 8 values8 Jul 2026Public8
GPT-Live-1 miniFirst card: 8 Jul 2026 · Public · 2 values8 Jul 2026Public2
GPT-5.6 LunaFirst card: 25 Jun 2026 · Public · 7 values25 Jun 2026Public7
GPT-5.6 SolFirst card: 25 Jun 2026 · Public · 93 values25 Jun 2026Public93
GPT-5.6 TerraFirst card: 25 Jun 2026 · Public · 5 values25 Jun 2026Public5
GPT-Rosalind-5.5First card: 3 Jun 2026 · Limited release · 2 values3 Jun 2026Limited release2
GPT-5.5 InstantFirst card: 4 May 2026 · Public · 15 values4 May 2026Public15
GPT-5.5First card: 23 Apr 2026 · Public · 42 values23 Apr 2026Public42
GPT-5.4 ThinkingFirst card: 5 Mar 2026 · Public · 29 values5 Mar 2026Public29
GPT-5.3 InstantFirst card: 3 Mar 2026 · Public · 10 values3 Mar 2026Public10
GPT-5.3-CodexFirst card: 5 Feb 2026 · Public · 30 values5 Feb 2026Public30
GPT-5.2-CodexFirst card: 18 Dec 2025 · Public · 13 values18 Dec 2025Public13
gpt-5.2-instantFirst card: 11 Dec 2025 · Public · 14 values11 Dec 2025Public14
gpt-5.2-thinkingFirst card: 11 Dec 2025 · Public · 41 values11 Dec 2025Public41
GPT-5.1-Codex-MaxFirst card: 18 Nov 2025 · Public · 39 values18 Nov 2025Public39
gpt-5.1-instantFirst card: 12 Nov 2025 · Public · 28 values12 Nov 2025Public28
gpt-5.1-thinkingFirst card: 12 Nov 2025 · Public · 40 values12 Nov 2025Public40
GPT-5-CodexFirst card: 15 Sep 2025 · Public · 23 values15 Sep 2025Public23
gpt-5-mainFirst card: 7 Aug 2025 · Public · 21 values7 Aug 2025Public21
gpt-5-thinkingFirst card: 7 Aug 2025 · Public · 68 values7 Aug 2025Public68
gpt-5-thinking-miniFirst card: 7 Aug 2025 · Public · 1 value7 Aug 2025Public1
gpt-5-thinking-nanoFirst card: 7 Aug 2025 · Public · 1 value7 Aug 2025Public1
gpt-oss-120bFirst card: 5 Aug 2025 · Open weights · 25 values5 Aug 2025Open weights25
gpt-oss-20bFirst card: 5 Aug 2025 · Open weights · 19 values5 Aug 2025Open weights19
o3First card: 16 Apr 2025 · Public · 50 values16 Apr 2025Public50
o4-miniFirst card: 16 Apr 2025 · Public · 38 values16 Apr 2025Public38
ChatGPT agentFirst card: not recorded · Public · 1 valueNot recordedPublic1
codex-1First card: not recorded · Public · 4 valuesNot recordedPublic4
GPT-4oFirst card: not recorded · Public · 28 valuesNot recordedPublic28
gpt-5-instantFirst card: not recorded · Public · 29 valuesNot recordedPublic29
gpt-5-thinking-helpful-onlyFirst card: not recorded · Internal only · 1 valueNot recordedInternal only1
GPT-5.1-CodexFirst card: not recorded · Public · 2 valuesNot recordedPublic2
GPT-5.2First card: not recorded · Public · 1 valueNot recordedPublic1
o1First card: not recorded · Public · 8 valuesNot recordedPublic8
o3-miniFirst card: not recorded · Public · 1 valueNot recordedPublic1