o1-mini
OpenAI
OpenAI's cost-efficient reasoning model optimized for STEM tasks. Uses same high-compute reinforcement learning pipeline as o1-preview but with smaller, faster model. Achieves 70% on AIME math competition (competitive with o1's 74.4%), placing it in top 500 US high school students. Reaches 1650 Elo on Codeforces (86th percentile of programmers), nearly matching o1's 1673 Elo. Generates answers 3-5x faster than o1-preview while being 80% cheaper. Excels at STEM reasoning but performs worse on non-STEM factual knowledge. Recommended upgrade to o3-mini available.
Strengths
- Nearly matches o1-preview performance on STEM tasks at 80% lower cost
- 70% on AIME math competition (top 500 US high school students)
- 1650 Elo on Codeforces competitive programming (86th percentile)
- 3-5x faster generation than o1-preview on word reasoning questions
- Excellent price-to-performance for math and coding tasks
- 128K context window for large problem contexts
- 65K max output tokens enables detailed solutions and explanations
Caveats
- Weaker on non-STEM factual knowledge (comparable to small LLMs like GPT-4o-mini)
- Not suitable for tasks requiring broad general knowledge or trivia
- No vision, tool use, or multimodal capabilities
- Slower than standard models due to reasoning process
- Newer o3-mini recommended as upgrade (higher intelligence, same latency/price)
- Overkill for simple tasks - use GPT-4o-mini for general queries
- Vendor
- OpenAI
- Context Window
- 128,000 tokens
- Max Output
- 65,536 tokens
- License
- Proprietary
- Release Date
- 2024-09-12
- Modalities
- text
| Benchmark | Score | Max |
|---|---|---|
| AIME 2024 | 70.0 | - |
| Codeforces | 1650.0 | - |
| livebench_language | 45.1 | 100.0 |
| livebench_coding | 48.1 | 100.0 |
| livebench_instruction_following | 72.6 | 100.0 |
Capabilities
Vision
Audio
Video
Tool Use
Resources
Pricing
$3.00 per 1M input tokens, $12.00 per 1M output tokens (80% cheaper than o1-preview)