OpenAI o1
Reasoning model trained with reinforcement learning to generate extended internal chains of thought before responding.
Best for: PhD-level STEM problems, cryptographic puzzles, and formal verification.
Specifications
- Provider: OpenAI
- Release date: September 2024
- Context window: 200,000 tokens
- Modalities: Text, Vision
- License: Proprietary Commercial API
- Pricing: $15.00 / MTok input, $60.00 / MTok output
- API available: Yes
Capabilities
- Internal chain-of-thought search
- Olympiad math reasoning
- Competitive programming
- Scientific reasoning