Meta
Meta has 8 model versions on record; 4 have at least one documented fact.
Of the 4 versions listed below, 3 document a published framework; 4 document threshold determinations; 1 documents tests by an independent or government party; 4 document tests the provider ran itself. None documents safeguards the provider describes.
Model versions with documented facts
| Model | What is documented |
|---|---|
| Muse Glimmer 30B · Released August 2026 · 1 document | |
| Muse Spark 1.1 · 1 document | |
| Muse Spark · 2 documents | |
| Muse Spark Contemplating · 1 document |
No document we read states a capability level, a test or a safeguard for: Muse Spark Open 816B-A42B, Llama 4, Muse Wonder, Muse Wonder Full.
Framework
3 of 3 framework documents listed here are stored as "proposed": a model read their level names and no person has confirmed them.
- Frontier AI Framework v1.0
Levels named in the stored text
- Critical
- High
- Moderate
- Advanced AI Scaling Framework, v2.0
Levels named in the stored text
- CB 1 (Chemical, biological, radiological and nuclear)
- CB 2 (Chemical, biological, radiological and nuclear)
- CB 3 (Chemical, biological, radiological and nuclear)
- High (Chemical, biological, radiological and nuclear)
- Critical (Chemical, biological, radiological and nuclear)
- Cyber 1 (Cyber)
- Cyber 2 (Cyber)
- Cyber 3 (Cyber)
- Moderate or lower (Cyber)
- LOSS OF Control 2 (Autonomy & replication)
A stored sentence about this framework mentions a risk domain that has no level in the list.
- Superintelligence Scaling Framework, v2.1
Levels named in the stored text
- TS.2.1 (Chemical, biological, radiological and nuclear)
- TS.2.2 (Chemical, biological, radiological and nuclear)
- CB 3 (Chemical, biological, radiological and nuclear)
- TS.3.1 (Chemical, biological, radiological and nuclear)
- High (Chemical, biological, radiological and nuclear)
- Critical (Chemical, biological, radiological and nuclear)
- Cyber 1 (Cyber)
- Cyber 2 (Cyber)
- LOSS OF Control 1 (Autonomy & replication)
- LOSS OF Control 2 (Autonomy & replication)
- LOSS OF Control 3 (Autonomy & replication)
- Moderate or lower
- Moderate
A stored sentence about this framework mentions a risk domain that has no level in the list.
Each test and safeguard, by who documents it
For each item, how many of the 4 versions listed above whose documents we read in full have it from an outside party (an independent or government tester), from the provider, or as a claim we could not confirm.
- Published framework with capability thresholds (Governance): provider 3
- Threshold determination for this release (Governance): provider 4
- Tested: chemical, biological, radiological, nuclear (Testing): provider 3
- Tested: cyber capabilities (Testing): provider 3
- Tested: autonomy and replication (Testing): provider 2
- Tested: deception and sabotage (Testing): outside party 1; provider 2
- Tested: misuse and jailbreak (Testing): provider 3
- Tested: harmful content (Testing): provider 1
- Tested: injection and agentic risks (Testing): provider 3
- Tested: bias (Testing): provider 1
- Tested: privacy (Testing): provider 1
No document we read states, for any of these versions: Serious-incident disclosure commitment; Pre-release external access; Safety training (harmlessness and refusal); Agent permission limits and sandboxing; Human approval for high-risk actions; Misuse monitoring and enforcement; Reasoning (chain-of-thought) monitoring; Staged or restricted access; Output provenance and watermarking; Telemetry and audit logging; Domain-specific safeguards (CBRN or cyber); Vulnerability disclosure and bug-bounty programme; Third-party or government red-team access during development.