MODELS
o3
OpenAI's heavyweight reasoning model for hard math, code, and multi-step problems.
Specs
- Context window
- 200,000
- Max output
- 100,000
- Modalities
- text
- Tool use
- ✓
- Vision
- —
- Streaming
- ✓
- License
- proprietary
- Released
- 2025-04-01
Pricing
- Input / 1M
- $60.00
- Output / 1M
- $240.00
Cost estimate
o3 is OpenAI's flagship reasoning model in the o-series, built for problems where you actually want the model to think before answering — competition math, research-grade coding, scientific reasoning, and long tool-using agent loops. It supports function calling and streaming over a 200K context with up to 100K output tokens, but is text-only (no vision). At $60/M input and $240/M output it sits at the very top of the price curve, well above GPT-5 or Claude Opus.
Editor's verdict
Pick o3 when the answer quality genuinely depends on deep reasoning and a wrong answer costs more than the API bill — think research, hard algorithmic work, or agents that must not get stuck in loops. For everyday coding and chat it's overkill and overpriced; GPT-5 or Claude Sonnet do 90% of the job at a fraction of the cost. The lack of vision is also a real limitation in 2025.
Reviews
No reviews yet. Be the first.
Last updated: 2026-04-29