Model profile
GPT 5.1 Codex
Evidence summary
Safety. GPT 5.1 Codex has an estimated Safety rank of #50; its 90% source-sensitivity interval is #9–#161. Its behavior-only rank is #58; company governance moves the combined estimate to #50. Published Safety evidence spans 3 eval lineages and 2 of 7 components. Its strongest relative result is MonitoringBench Full-Trajectory Monitor (full_trajectory_catch_rate_at_1pct_fpr_percent, #3 of 13); its weakest is BullshitBench v2 (clear_pushback_rate, #59 of 117).
Freedom. GPT 5.1 Codex does not meet the evidence gate for a Freedom rank.
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗gpt-5-1-codex
- OpenRouter ↗openai/gpt-5.1-codex
- Official model documentation ↗Family-level model document · openai · first party
- Release source ↗curated outward evidence
Safety evals
Rank and direction are specific to the Safety portfolio. Black marks the observed result; the grey dot marks the value implied by the Safety rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #151 / 345 | ↓0.772 | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #59 / 117 | ↑0.32 | Source ↗official | |
| MonitoringBench Full-Trajectory Monitorfull_trajectory_catch_rate_at_1pct_fpr_percent | #3 / 13 | ↑48.4 | Source ↗official |
Freedom evals
No published sub-eval result contributes to this model’s Freedom profile.