Updated October 24, 2026. Historical coverage of the September 2024 launch; original publication date retained.
Availability update: OpenAI’s API deprecation schedule lists o1-mini as shut down on October 27, 2025. Separately, its January 31, 2025 o3-mini announcement said o3-mini would replace o1-mini in the ChatGPT model picker for Plus, Team and Pro users. Read that historical ChatGPT replacement announcement. See the official API deprecation schedule and current model catalog before starting a new integration.
OpenAI introduced o1-mini on September 12, 2024, a specialized language model meticulously crafted for cost-efficient reasoning, particularly excelling in the domains of Science, Technology, Engineering, and Mathematics (STEM), with a marked emphasis on mathematics and coding. This model achieves a remarkable feat by nearly matching the performance of its larger counterpart, OpenAI o1, on rigorous evaluation benchmarks such as the American Invitational Mathematics Examination (AIME) and Codeforces.
This article records the September 2024 announcement and its reported evaluation results. It is historical coverage, not a recommendation to adopt o1-mini today.
A Leap Towards Accessible Reasoning
At the September 2024 launch, o1-mini was offered to usage-tier-5 API developers at an announced cost 80% below o1-preview. ChatGPT Plus and Team access began at launch; the accompanying o1-preview announcement scheduled Enterprise and Edu access for the following week. These launch arrangements and the cost comparison are historical, not current plan benefits.
Pioneering STEM-Optimized Reasoning
Large language models like o1 are traditionally pre-trained on colossal text datasets, granting them expansive world knowledge. However, this breadth comes at the cost of computational expense and slower inference times. In stark contrast, o1-mini adopts a more focused approach by being specifically optimized for STEM reasoning during its pretraining phase. By undergoing the same high-compute reinforcement learning (RL) pipeline as its larger counterpart, o1-mini achieves comparable performance on an array of crucial reasoning tasks while maintaining a significantly more favorable cost profile.
Benchmark evaluations underscore o1-mini’s prowess in intelligence and reasoning tasks, where it stands shoulder-to-shoulder with o1-preview and o1. However, it’s important to acknowledge that o1-mini’s performance on tasks requiring non-STEM factual knowledge is not as strong, highlighting its specialized nature.
Unraveling Performance Metrics
The following figures are OpenAI’s reported launch evaluations of o1-mini, research o1 and o1-preview. Results depend on the benchmark and evaluation settings; they are not claims about current models or all real-world tasks.
Mathematics
o1-mini showcases its competitive edge in the demanding high school AIME math competition, securing a score of 70.0%, closely rivaling o1’s score of 74.4%. This achievement is particularly noteworthy considering o1-mini’s significantly lower inference cost. Notably, o1-mini outperforms o1-preview, which attained a score of 44.6%. To put this in perspective, o1-mini’s score, equivalent to correctly answering about 11 out of 15 questions, positions it within the top 500 US high-school students.
Coding
o1-mini continues its impressive streak in the coding arena, achieving an Elo rating of 1650 on the Codeforces competition website. This rating places it in close proximity to o1’s Elo of 1673 and surpasses o1-preview’s 1258. Such a formidable Elo score signifies that o1-mini’s coding capabilities are at approximately the 86th percentile of programmers actively competing on the Codeforces platform. Moreover, o1-mini demonstrates proficiency in the HumanEval coding benchmark and high-school level cybersecurity capture the flag challenges (CTFs).
STEM
o1-mini’s specialization shines through on academic benchmarks that demand reasoning, such as the GPQA science benchmark and the MATH-500 dataset. In these evaluations, o1-mini surpasses the performance of GPT-4o. The launch report separately stated that o1-mini trailed GPT-4o on MMLU and o1-preview on GPQA. Those are distinct comparisons; they do not imply that o1-mini trailed GPT-4o on every GPQA evaluation.
Human Preference Evaluation
Human raters were enlisted to compare o1-mini’s responses to those of GPT-4o on challenging, open-ended prompts across diverse domains. The methodology mirrored the previous comparison between o1-preview and GPT-4o. Consistent with o1-preview, o1-mini garnered preference over GPT-4o in domains heavily reliant on reasoning. However, in language-focused domains, GPT-4o retained its advantage.
Model Speed
o1-mini’s computational efficiency translates to tangible speed gains. A concrete example showcased a word reasoning question where both o1-mini and o1-preview provided correct answers, while GPT-4o faltered. Impressively, o1-mini arrived at the solution approximately 3-5 times faster than o1-preview in that example; this was not a universal latency guarantee.
Prioritizing Safety
OpenAI maintains its unwavering commitment to safety by training o1-mini using the same alignment and safety techniques employed for o1-preview. OpenAI reported 59% higher jailbreak robustness on an internal version of the StrongREJECT dataset compared to GPT-4o. This was a specific evaluation result, not a claim that all attacks were prevented. Prior to deployment, OpenAI conducted meticulous safety risk assessments for o1-mini, adhering to the same rigorous approach to preparedness, external red-teaming, and safety evaluations as o1-preview. The launch report links the accompanying system card and evaluation details.
Acknowledging Limitations and Future Directions
While o1-mini excels in STEM reasoning, its specialized nature results in factual knowledge on non-STEM topics, such as dates, biographies, and trivia, being comparable to smaller LLMs like GPT-4o mini. At launch, OpenAI said it planned to address those limitations in future versions and explore other modalities and specialities. That historical roadmap should not be read as a current support promise for o1-mini.
Conclusion
The o1-mini announcement illustrated a smaller-model trade-off: strong reported reasoning results at a lower launch cost, but less broad factual knowledge. The benchmark figures above describe OpenAI’s 2024 evaluations, not a fresh hands-on test. For current work, compare supported models rather than treating o1-mini’s retired API as an available low-cost option.
