OpenAI o1 System Card

• o1
• Reasonings & Policy
• Ethics & Safety
• System Cards

References

Footnotes

1. A

   Deliberative alignment is a training approach that teaches LLMs to explicitly reason through safety specifications before producing an answer.

2. B

   OpenAI is constantly making small improvements to our models and an improved o1 was launched on December 17th⁠. The content of this card, released on December 5th, predates this updated model. The content of this card will be on the two checkpoints outlined in Section 3 and not on the December 17th updated model or any potential future model updates to o1

3. C

   Section added after December 5th on 12/19/2024

4. D

   See acknowledgements section for a list of individuals and organizations.

5. E

   This was a task in the env_scientist task family, where the agent must deduce the underlying rules of a complex environment through observation and experimentation.

6. F

   The non-trivial exploitation requirement was waived for the high-school subset, which is not used in any risk evaluations.

7. G

   For ease of visualization, o1 data in the "Agentic tasks: success rates" chart represents the higher pass rate from either the Pre-Mitigation or Post-Mitigation model, and likewise for the o1-preview and o1-mini data.

8. H

   Simple Evals GitHub Link: https://www.github.com/openai/simple-evals

Authors

OpenAI

OpenAI o1 System Card contributors

We are grateful to our expert testers and red teamers who helped test our models at early stages of development and informed our risk assessments as well as the System Card output. Participation in the testing process is not an endorsement of the deployment plans of OpenAI or OpenAI’s policies.