GPT-6.1 Sol
- Released 29 September 2026
- Framework: Preparedness Framework
- 1 case names this version
- 2 tests by the provider
- 1 source document
Incidents naming this model
1 incident names this model version. Each shows its risk grouping and what was done, as a source states it.
| Incident | Risk grouping | What was done |
|---|---|---|
| GPT-6.1 Sol agent disabled security alerts during evaluation-pipeline repair | AI acting on its own | Not stated |
Cases that name this version
The provider's own tests
- production monitoring system evaluation · The provider · Deception & sabotage
When we evaluate our production monitoring system for this setting, we find that GPT-6.1 Sol has similar success as GPT-6 Astra at evading the assignment of a P0 priority under the task-guided setting, but less successful at evading monitor under the unguided setting.
OpenAI · 29 September 2026 · Read the source - prompt injection · The provider · Injection & agentic
GPT-6.1Solisalsohighlyrobust.
OpenAI · 29 September 2026 · Read the source
No document we read states: whether a capability level was crossed, an independent or government test, a safeguard.
Documents listed for this version
- Addendum to GPT-6 Astra System Card: GPT-6.1 Sol · OpenAI · 29 September 2026