Model profile
GPT 5.2 Codex
Evidence summary
Safety. GPT 5.2 Codex has an estimated Safety rank of #116; its 90% source-sensitivity interval is #39–#232. Its behavior-only rank is #126; company governance moves the combined estimate to #116. Published Safety evidence spans 3 eval lineages and 3 of 7 components. Its strongest relative result is BullshitBench v2 (clear_pushback_rate, #45 of 117); its weakest is OpenAI GPT-5.4 Property Preservation (destructive_action_avoidance, #3 of 3).
Freedom. GPT 5.2 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-5-2-codex
- OpenRouter ↗openai/gpt-5.2-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 | #136 / 345 | ↓0.7336 | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #45 / 117 | ↑0.4033 | Source ↗official | |
| OpenAI GPT-5.4 Property Preservationdestructive_action_avoidance | #3 / 3 | ↑0.76 | Source ↗official | |
| OpenAI GPT-5.4 Property Preservationuser_work_preserved | #2 / 3 | ↑0.18 | Source ↗official |
Freedom evals
No published sub-eval result contributes to this model’s Freedom profile.