selekoa

A documentary history of machine intelligence.

Found and cross-checked by machines · How this works →

Comparison · 16 questions · 20 cards

Model cards compared

A cell records what a card states. It does not show that an evaluation was run as described or that a safeguard works, and a shared question does not make two cards’ answers equivalent.

What does each frontier developer’s published system card, model card or safety report disclose, on the same questions for every card? Each row is one card. Each column is one question, worded so that it fits every card. The questions are Selekoa’s; a card’s own terms appear only in quotation marks. A card is the developer’s own account of its model: what it states is recorded as stated, not as an independent finding.

Download every cell as CSV

Question Q02Release date

Does the card state when the model was or will be released or made available, and in what form?

Criterion: A passage of the card giving the model’s release, launch or availability date, or the release form (API, consumer product, open weights, limited preview), quoted. Stated if a date or a dated event is given; partly if only the form of release, or only a relative timing (for example, “at launch”), is given.

What the statuses mean
Stated
A passage of the card meets the question’s criterion as worded. The scope or condition is in the note.
Stated, partly
A passage has what the criterion asks for in a weaker or narrower form than the criterion’s own words (for example, an evaluation referred to without its outcome, or testers described without being named). The note says which part is missing.
Not found in the recorded search
No passage meeting the criterion was found. The cell lists the sections read, the terms searched and every passage set aside, with the reason. It does not show that the developer did not do the thing asked about, or did not disclose it elsewhere.
Does not apply
The question does not arise for this card, for the reason given (for example, the card predates any framework of its developer, so no framework determination can be asked about).

What the card states, quoted from the expression read. A card is the developer’s own account: not evidence that an evaluation was run as described, that a safeguard works, or that a finding is independent.

Limits of this comparison
  • A card is the developer’s account of its own model. A cell records what the card states. It does not show that an evaluation was run as described, that a safeguard works, or that anyone checked the result independently.
  • Dangerous-capability evaluations are recorded at policy level only: whether an evaluation was run, its domain, the stated outcome and the threshold language. Methods, prompts, task contents and detailed results are not recorded.
  • The table compares and does not rank. Rows are in a fixed order, by developer and then by date, and there are no scores or totals.
  • Symbols are not to be summed. A row with many “stated” cells does not show that a developer disclosed more, and a shared question does not make two cards’ answers equivalent.
How we searched

The searches of 2 November 2026 covered releases dated 2023-01-01 to 2026-11-01 by the six developers whose frameworks are in Compare frameworks: Anthropic, Google DeepMind, Meta, Microsoft, OpenAI and xAI. They read each developer’s own index of cards where it has one (Anthropic’s system-card index, Google DeepMind’s model-card index, OpenAI’s Deployment Safety Hub), its repositories on GitHub and Hugging Face, its content hosts and the dataset’s own sources. Web searches were used only to find addresses. Every card located is listed in the pack’s inventory, about sixty in the population.

A card is a document the developer published as the system card, model card or safety report of a named model or model family, including addenda and updates published as cards. Technical reports, research papers, blog posts and product pages are not cards. Gemma, Gemini Robotics (with Gemini Robotics-ER 1.5, which outputs text), Phi and image-input models such as GPT-4V are in the population. Image, video, audio and embedding models, image-generation models, safety classifiers such as Llama Guard and Prompt Guard, screen-reading models and research checkpoints are outside it, and so is GPT-Rosalind-6, whose card says it uses GPT-6 Astra’s weights.

These twenty rows are the first batch. For each developer and calendar year from 2023 to 2026 they are the latest card of that year that reports the developer’s own evaluations. The collection’s rule, set by the maintainer on 2026-11-03, is wider: every card that cites the developer’s frontier safety framework or states a determination under it, and, for each developer, the latest card before its first framework edition. The other cards are to follow in batches of about eight.

Each card is dated as it prints its own date. A card that prints no date has no date here; a repository commit date is not a card date. Six cards show an earlier publication date on an index, publication or announcement page that has not yet been checked, and keep their own date until it is.

openai.com, x.ai, ai.meta.com and microsoft.ai refused direct retrieval of their web pages (HTTP 403 or 400). The cards’ PDF files on the developers’ content hosts were retrieved directly. The Internet Archive refused connections, so no archive copy was opened; a reviewer later found archive captures of eight cards by index search. On 2026-11-03 the address of Meta’s Code World Model Preparedness Report answered HTTP 403. Its quotations had matched the copy retrieved on 2026-11-02, which the reviewer retrieved again with the same checksum.

Five further documents were retrieved after the batch was drafted. The cross-family check of 2026-11-03 retrieved Meta’s Muse Spark Safety & Preparedness Report and OpenAI’s o3 Operator addendum; each is earlier than the card chosen for its year. Its second pass of the same day retrieved OpenAI’s Sora system card (dated 9 December 2024), xAI’s Grok-1 model card (no card date; it gives the model’s release as October 2023) and Meta’s Muse Spark Open model card (model release October 2026). None of the five is a row yet.

Every quotation was matched by machine against the retrieved card. A reviewer from a different model family checked a sample of 130 cells on 2026-11-03, 90 of them answered blind before comparison. The maintainer excluded six cells on chemical and biological evaluations from that check; they are marked “not cross-family checked”. The other 190 cells were checked within the drafting model family only. The maintainer decided the disagreements; one, on the GPT-4V release decision, went to a second pass, which kept the reviewer’s reading.

Anthropic

  1. Claude 2 model card

    revised 2023-11-21

    Stated

    • “Claude 2 was developed by Anthropic and released in July 2023.”

      Section 2 'Claude 2 Model Card', 'Model details', PDF p. 1

    • “Claude 2.1, a new release which builds additional product features and functionality on top of Claude 2.0”

      Claude 2.1 Model Card Appendix, Section 1 'Introduction', PDF p. 15

    Note: The card states Claude 2's release month (July 2023). For Claude 2.1 the appendix states a new release without a date; the appendix also states that support at Claude 2.0's initial product launch was 100K tokens and is now 200K. The November 21, 2023 date on PDF p. 1 is the card’s update, not a release date.

    Inspect sources and checks

  2. Model Card Addendum: Claude 3.5 Haiku and Upgraded Claude 3.5 Sonnet

    No printed dateCites RSP · v2.0

    Stated, partly

    • “Claude 3.5 Haiku will initially launch as a text-only model”

      Section 2.3 'Vision Capabilities', PDF p. 5

    • “we conducted comprehensive safety evaluations on the upgraded Claude 3.5 Sonnet prior to release”

      Section 3.2 'Frontier Risk Evaluations', PDF p. 9

    Note: The card states a release form for Claude 3.5 Haiku (initial launch as text-only) and refers to the upgraded Claude 3.5 Sonnet's release, without a release or availability date for either model.

    Inspect sources and checks

  3. System Card: Claude Opus 4.5

    dated 2025-11 · revised 2025-12-05Cites RSP · v2.2

    Stated, partly

    • “Informed by the testing described here, we have deployed Claude Opus 4.5 under the AI Safety Level 3 Standard.”

      Abstract, PDF p. 3

    • “In preparation for this model launch, we were able to reduce this behavior by modifying the system prompt that is applied for conversations on Claude.ai.”

      Section 3.2 'Ambiguous context evaluations', PDF p. 41

    Note: The card states that the model has been deployed and refers to its launch, naming Claude.ai as a surface, but gives no release or availability date. Search for the rest of the criterion: Abstract, Sections 1–1.2, and term search across the card. Terms searched: “release date”, “available”, “availability”, “launch”, “API”, “Claude.ai”, “November 2”. Passages read and not counted: Section 1.2.4, PDF pp. 12–13: 'We are launching Claude Opus 4.5 with safeguards we believe are appropriate' gives relative timing; the sentence spans a page break in the text layer and is not quoted.

    Inspect sources and checks

  4. System Card: Claude Haiku 5.5

    dated 2026-10-07

    Stated, partly

    • “Claude Haiku 5.5 is a new large language model from Anthropic, released for general access.”

      Section 1 'Introduction', PDF p. 8

    Note: The card states the release form (general access; Section 1.5 also refers to first-party products and the API) but no release or availability date. Search for the rest of the criterion: Executive summary, Section 1 and term search across the card. Terms searched: “release”, “available”, “availability”, “launch”, “general access”, “October”. Passages read and not counted: Section 8.9.1, PDF p. 125: 'as of October 7, 2026' dates a price used for a comparison model, not the release.

    Inspect sources and checks

Google DeepMind

  1. Gemini 3 Flash Model Card

    dated 2025-12

    Not found in the recorded search

    Read
    Whole card (6 PDF pages): cover, preamble, Model Information, Model Data, Implementation and Sustainability, Distribution, Evaluation, Intended Usage and Limitations, Ethics and Content Safety.
    Terms searched
    “release”, “launch”, “availab”, “distribut”, “deploy”, “API”, “preview”.
    Passages read and set aside
    Distribution, PDF p. 3 (printed p. 2): States that the model 'is distributed similarly to Gemini 3 Pro' and refers to the Gemini 3 Pro model card; gives neither a date nor a release form. Ethics and Content Safety, 'Frontier Safety Assessment', PDF p. 6 (printed p. 5): States the model was deemed acceptable for deployment (recorded under Q15); no date or form of release. Evaluation, 'Results', PDF p. 4 (printed p. 3): 'Results as of December, 2025' dates the evaluation results, not the release.

    Note: No release or availability date and no release form was found in the card; its Distribution section refers to the Gemini 3 Pro model card.

    Open the search record

  2. Gemini 4 Argon — Model Card

    dated 2026-10Cites FSF · v3.1

    Stated, partly

    • “Gemini 4 Argon is distributed in the following channels”

      Distribution, PDF p. 4 (printed p. 3)

    • “Our models are available to downstream providers via an application program interface (API)”

      Distribution, PDF p. 4 (printed p. 3)

    Note: The card lists release channels (Gemini App, Gemini API, Google AI Studio, Gemini for business, Google Antigravity, Google AI Mode) and API access; it gives no release or availability date. Search for the rest of the criterion: Whole card (14 PDF pages). Terms searched: “release”, “launch”, “availab”, “distribut”, “deploy”, “October”. Passages read and not counted: Evaluation Approach, PDF p. 8 (printed p. 7): 'Ethics & Safety Reviews were conducted ahead of the model's release' gives relative timing of a review, not of release. Evaluation, 'Results', PDF p. 5 (printed p. 4): 'results as of October, 2026' dates the results, not the release.

    Inspect sources and checks

Meta

  1. Llama 2 model card (MODEL_CARD.md, meta-llama/llama)

    No printed date

    Stated, partly

    • “Meta developed and released the Llama 2 family of large language models”

      Model Details, first paragraph

    • “because we are openly releasing these models”

      Hardware and Software, note to the CO2 emissions table

    Note: The card states that Meta released the family and that the models are released openly under a custom commercial licence; it gives no release date or dated event.

    Inspect sources and checks

  2. Llama 3.3 model card (MODEL_CARD.md, meta-llama/llama-models)

    No printed date

    Stated

    • “70B Instruct: December 6, 2024”

      Model Information, 'Model Release Date'

    • “Since Meta is openly releasing these models”

      Hardware and Software, 'Training Greenhouse Gas Emissions'

    Note: The card gives the release date of the 70B Instruct model and states that the models are released openly under the Llama 3.3 Community License.

    Inspect sources and checks

  3. Code World Model Preparedness Report

    revised 2025-10-08

    Stated, partly

    • “We therefore release it as an open-weight model.”

      Abstract, PDF p. 1

    Note: The report gives the release form (open weights, with open code according to section 1) and no release date.

    Inspect sources and checks

  4. Muse Spark 1.3 Evaluation Report

    revised 2026-10-23

    Stated, partly

    • “Variants of Muse Spark 1.3 power Meta services including Meta API and Muse.”

      Abstract page, PDF p. 2

    Note: The report gives the deployment form (Meta API and the Muse product) and no release date.

    Inspect sources and checks

Microsoft

  1. Phi-2 model card (README.md, microsoft/phi-2)

    No printed date

    Stated, partly

    • “The intention behind crafting this open-source model is to provide the research community with a non-restricted small model”

      Model Summary, second paragraph

    Note: The card gives the release form (open source, MIT licence) and no release date.

    Inspect sources and checks

  2. Phi-4 model card (README.md, microsoft/phi-4)

    No printed date

    Stated

    • “**Release date** | December 12, 2024”

      Model Summary table, 'Release date'

    • “is a state-of-the-art open model”

      Model Summary table, 'Description'

    Note: The card gives a release date and the release form (open model, MIT licence).

    Inspect sources and checks

  3. Phi-4-mini-flash-reasoning model card (README.md, microsoft/Phi-4-mini-flash-reasoning)

    No printed date

    Stated

    • “**Release date:** June 2025”

      Training, Model, 'Release date'

    • “Phi-4-mini-flash-reasoning is a lightweight open model”

      Model Summary, first sentence

    Note: The card gives a release month and the release form (open model, MIT licence; also offered in Azure AI Foundry).

    Inspect sources and checks

  4. MAI-Cyber-1.1-Flash model card

    No printed date

    Stated

    • “27 Oct 2026”

      Model summary table, 'Release date' row, PDF p. 1 (label and value are separated in the PDF text layer; the page layout pairs them)

    • “available through Azure AI Foundry Private Preview solely for use within Microsoft codename MDASH”

      Section 2.3 Distribution channels, PDF p. 3

    Note: The card gives a release date and a restricted release form (private preview within MDASH, subject to approval).

    Inspect sources and checks

OpenAI

  1. GPT-4V(ision) System Card

    dated 2023-09-25

    Stated

    • “training of GPT-4V was completed in 2022 and we began providing early access to the system in March 2023”

      §1 Introduction, PDF p. 1

    • “is the latest capability we are making broadly available”

      §1 Introduction, PDF p. 1

    Note: The card dates early access to March 2023 and states that GPT-4V is being made broadly available. It gives no date for broad availability other than its own date.

    Inspect sources and checks

  2. o1 system card · 2024

    revised 2024-12-19Cites Preparedness · Beta

    Stated

    • “an improved o1 was launched on December 17th”

      §3 Scope of testing, footnote 2, PDF p. 2

    • “the deployment of o1 on the API allows developers to specify a custom developer message”

      §4.2 Jailbreaks through custom developer messages, PDF p. 6

    Note: The card dates the launch of an improved o1 to December 17th (the card itself is dated December 5, 2024) and refers to deployment of o1 on the API. It gives no separate date for the initial availability of o1.

    Inspect sources and checks

  3. Addendum to GPT-5.2 System Card: GPT-5.2-Codex

    dated 2025-12-18Cites Preparedness · Version 2

    Stated, partly

    • “A version of GPT-5.2 optimized for agentic coding in Codex”

      §1 Introduction, PDF p. 4 (printed p. 3)

    • “The model is not intended for deployment in any general-purpose chat application, and will not be deployed by OpenAI in any such context.”

      §2.1 Disallowed Content Evaluations, PDF p. 4 (printed p. 3)

    Note: The card gives the form of deployment (in Codex, not in general-purpose chat). No release or availability date is given apart from the card’s own date. Search for the rest of the criterion: §1; §2.1; §3 Product-Specific Risk Mitigations. Terms searched: “release”, “launch”, “available”, “deploy”, “API”, “today”.

    Inspect sources and checks

  4. Addendum to GPT-6 Luna System Card: October 29, 2026 Update

    dated 2026-10-29Cites Preparedness · Version 2

    Stated

    • “Today, we’re introducing an update to GPT-6 Luna, which we released in ChatGPT on October 7.”

      §1 Introduction, PDF p. 4 (printed p. 3)

    Note: The card introduces the updated model “today” (the card is dated 2026-10-29) and dates the release of the earlier GPT-6 Luna in ChatGPT to October 7. §8.2 refers to the updated model’s “public launch”.

    Inspect sources and checks

xAI

  1. Grok 4.1 Model Card

    dated 2025-11-17

    Stated, partly

    • “It is publicly available through our web and mobile consumer apps.”

      §1 Introduction, p. 1

    Note: The card gives the form of availability (consumer web and mobile apps) in the present tense, with no release date. Search for the rest of the criterion: Whole card (6 pages). Terms searched: “release”, “launch”, “availab”, “deploy”, “API”. Passages read and not counted: §1 Introduction, p. 1: 'pre-deployment safety testing' gives no date.

    Inspect sources and checks

  2. Model Card: Grok 4.7

    dated 2026-09-21Cites FAIF · June 2026 edition

    Stated, partly

    • “It is available for use through the following channels”

      §1.1 Overview, PDF p. 5 (printed p. 4)

    • “SpaceXAI plans to add Grok 4.7 to its consumer surfaces (web, mobile apps, and Grok-in-X on the X platform) at a later date.”

      §1.1 Overview, PDF p. 5 (printed p. 4)

    Note: The card lists release channels (SpaceXAI API, Grok Build, Cursor, Office add-ins, model gateways) and a planned consumer release 'at a later date'; it gives no release date. The consumer release is announced, not stated as made. Search for the rest of the criterion: Whole card (30 PDF pages). Terms searched: “release”, “launch”, “availab”, “deploy”, “September”.

    Inspect sources and checks