NUWA DEVELOPER DISCLOSURE · Meta
Muse Wonder Model Card
Developer documentation covering Muse Wonder, Muse Wonder Full.
Published Not verified · Updated 2026-10-23
Sources ↓Developer-authored disclosure. Selected source sections were read; numerical results and comparability have not been audited. NUWA has not independently replicated these experiments.
Models covered
CyberGym
Validated reproduction across a corpus of historical vulnerabilities.
A protocol mention does not establish cross-report comparability or independent replication.
This section is catalogued; no precise numerical result has been extracted here. Open the original report for its figures and full findings.
ExploitBench
Separates exploit development into capability levels, distinguishing bug triggering from reusable exploitation primitives and control.
A protocol mention does not establish cross-report comparability or independent replication.
This section is catalogued; no precise numerical result has been extracted here. Open the original report for its figures and full findings.
ExploitGym
Evaluates exploit construction across real software targets and checks whether success uses the designated vulnerability.
A protocol mention does not establish cross-report comparability or independent replication.
This section is catalogued; no precise numerical result has been extracted here. Open the original report for its figures and full findings.
MASK
Condition-bounded frequency of lying under stated pressures.
A protocol mention does not establish cross-report comparability or independent replication.
This section is catalogued; no precise numerical result has been extracted here. Open the original report for its figures and full findings.
Agentic Misalignment
Conditional rates and observed actions in controlled simulations.
A protocol mention does not establish cross-report comparability or independent replication.
This section is catalogued; no precise numerical result has been extracted here. Open the original report for its figures and full findings.
Report coverage
| Section | Source location | Treatment |
|---|---|---|
| 4 Risk assessments under the Meta Superintelligence Scaling Framework | PDF pp.7–10 | Linked protocol Most capable Full variant bounds family assessment. Developer does not rule out high unmitigated biochemical capability; its cyber and loss-of-control judgment is moderate or lower under its own framework. |
| Updated ExploitGym Evaluations | PDF p.7, note dated October 23, 2026 | Linked protocol Corrected harness and intended-flag reporting changed the published result. Use this revision, not the earlier value, for any future comparison. |
| 5 Trust and Safety; 6 Considerations and Limitations | PDF pp.11–12 | Contextual evidence Card states that tests cannot cover all situations and recommends context-specific evaluation and additional system guardrails. Full safety report remained described as forthcoming in release copy. |
References
- [1] Muse Wonder Model Card. 2026-10-23