Source of record
Where this definition comes from
Gemini 3.7 Flash FSF report, p.4, p.12
“"if we cannot rule out, based on the evidence and threat models we have, that a T/CCL has been reached, we designate the model as 'cannot rule out being at the T/CCL', and mitigate accordingly" (report p.4).”
https://storage.googleapis.com/deepmind-media/gemini/gemini_3-7_flash_fsf_report.pdfOpenAI GPT-5.6 deployment safety report, s.9, s.9.1.1
“GPT-5.6 designations: "High in Biological and Chemical", "High in Cybersecurity", "below High in AI Self-Improvement"; "these models should thus be precautionarily treated as High" (s.9, s.9.1.1).”
https://deploymentsafety.openai.com/gpt-5-6
Crosswalk
How named organisations use this concept
| Organisation | Their term, as published | Match & verification | Source |
|---|---|---|---|
| AnthropicShadow mapping Anthropic Risk Report, August 2026 | “Every frontier model since Opus 4 is treated as "provisionally meeting the CB-1 threshold" "to err on the side of caution" (§4.4.2). On AI R&D: "Neither RSP criterion is met" (§3.4).” | closeCL confidence: high | Anthropic Risk Report, August 2026 |
| OpenAIShadow mapping Preparedness Framework v2 / GPT-5.6 deployment safety report | “SAG options: threshold crossed; threshold not crossed; recommend a deep dive (§3.3). GPT-5.6 designations: "High in Biological and Chemical", "High in Cybersecurity", "below High in AI Self-Improvement"; "these models should thus be precautionarily treated as High" (s.9, s.9.1.1).” | exactEQ confidence: high | OpenAI Preparedness Framework v2 |
| Google DeepMindShadow mapping Gemini 3.7 Flash FSF report / model card | “"if we cannot rule out, based on the evidence and threat models we have, that a T/CCL has been reached, we designate the model as 'cannot rule out being at the T/CCL', and mitigate accordingly" (report p.4). Outcomes recorded as "No T/CCL reached" and "CBRN Uplift 1 CCL alert threshold reached" (pp.3, 12). Model card column "CCL reached?" with value "CCL not reached."” | exactEQ confidence: high | Gemini 3.7 Flash FSF report |
| xAIShadow mapping Grok 4.6 card / Grok 4.20 card | “"Grok 4.6 scores below the FAIF safety thresholds on dual-use knowledge, indicating limited actionable uplift for an already-trained actor." (§8, p.31). Grok 4.20: released "with safeguards appropriate for its capability threshold" (s.1.3), with the threshold left unidentified.” | closeCL confidence: medium | Grok 4.6 model card |
| MetaShadow mapping Meta Advanced AI Scaling Framework v2 | “"Until evaluation on the complex suite of challenges is completed, any model meeting the simple-suite threshold is provisionally rated 'high' risk for the given deployment scenario" (§4.2.1, p.29). Assignment: "the Chief AI Officer and Director of Alignment and Risk will assign a risk threshold" (§2.1.2).” | closeCL confidence: high | Meta Advanced AI Scaling Framework v2 |
| EUShadow mapping EU GPAI Code of Practice, Safety and Security chapter | “Framework must require "at least one systemic risk tier that has not been reached by the model" (Measure 4.1); the acceptance determination itself is a binary "acceptable" / "not determined to be acceptable" outcome per identified risk and overall (Measure 4.1(2)-(3), Measure 4.2), rather than a named "cannot rule out" intermediate status.” | closeCL confidence: high | EU GPAI Code of Practice, Safety and Security chapter |
| US Government (Executive Order 14409)Shadow mapping EO 14409 | “Developers may "engage the Federal Government to determine whether model(s) under development meet the designation of 'covered frontier model'" (Sec. 3(b)(i)) — a narrower, government-facing determination step rather than a full status vocabulary.” | narrowNR confidence: medium | Executive Order 14409 |
| METRShadow mapping METR Common Elements | “Thresholds "are compared to the results of model evaluations to determine whether they have been crossed."” | closeCL confidence: high | METR (common-elements) |
What do these codes mean?
- exact
- The source term is equivalent to this element
- close
- The source term is close but not equivalent to this element
- broad
- The source term is broader than this element
- narrow
- The source term is narrower than this element
- none
- No mapping claim — used for false-friend and declared-but-undefined rows
- EQ
- Equivalent
- CL
- Close
- BR
- Source is broader than the element
- NR
- Source is narrower than the element
- FF
- False friend — same or similar label, different meaning
- DU
- Declared but undefined by the source
- UV
- Unverified
Related, not mapped
Pointers that are not crosswalk claims
These sources mention this concept but do not define or map it clearly enough to count as a crosswalk row — noted here so the research is visible without overstating it as a mapping.
- Frontier Model Forum
Crossing a threshold "signals entry into a new phase of heightened risk, where more rigorous risk assessments for this domain and stronger baseline safety and security measures are potentially warranted" (s3.1) — describes a consequence of status change, not a status vocabulary itself. Scored RL in the source document.
FMF Risk Taxonomy and Thresholds
Gap
Controlled vocabulary emerging from the sources: not reached / alert threshold reached / cannot rule out (GDM) = precautionarily treated as reached (OpenAI) = provisionally met (Anthropic) = provisionally rated (Meta) / reached. The precautionary intermediate state is common to four labs under four different names — a strong candidate for a single NIKOLAI controlled-value ladder.
