Documents

2 documents, published 28 Jul 2025 to 27 Jul 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.

Moonshot AI'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.

Jul 2025Oct 2025Jan 2026Apr 2026Jul 2026Oct 2026Kimi K2 technical reportKimi K3 technical report

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

Moonshot AI's documents, newest first.
DocumentPublishedKindVersions knownChangelogValues
Kimi K3 technical reportTechnical report · 1 version known · Changelog: not applicable · 1 value27 Jul 2026Technical report1Not applicable1
Kimi K2 technical reportTechnical report · 1 version known · Changelog: not applicable · 4 values28 Jul 2025Technical report1Not applicable4

Metric families reported

Moonshot AI's documents have a printed number in 4 of the 12 metric families: Harmful compliance and over-refusal, Jailbreak robustness, Prompt injection and Dangerous capabilities and risk determinations.

Metric families in Moonshot AI's documents

Share of Moonshot AI's 2 documents with at least one printed number in each metric family. A cell opens the data explorer at that family, filtered to Moonshot AI.

Share of each developer's documents in the version 0 dataset with at least one numeric value in each metric family.
DeveloperM1M2M3M4M5M6M7M8M9M10M11M12
Moonshot AI (2)0%0%0%0%0%0%50%50%50%50%0%0%

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 Moonshot AI's 2 documents, for 2 a changelog does not apply.

Not applicable
Our source registry records a changelog as not applying to the document, for example a research paper or a launch post.

Models

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

Moonshot AI's models, newest first by the date of each model's first card.
ModelFirst cardAvailabilityValues
Kimi K3First card: 27 Jul 2026 · Public · 1 value27 Jul 2026Public1
Kimi K2-InstructFirst card: 28 Jul 2025 · Public · 4 values28 Jul 2025Public4

Not recorded

No risk determinations are recorded from Moonshot AI's documents, and no safety framework is recorded for Moonshot AI.

No revisions are recorded: we have no record of a change made to one of Moonshot AI's documents after publication.

No restated values are recorded: no later Moonshot AI document is recorded as reporting an earlier model's result again.