MODELS
o3-mini
Cheap reasoning model tuned for STEM, code, and structured tool calls.
Specs
- Context window
- 200,000
- Max output
- 65,536
- Modalities
- text
- Tool use
- ✓
- Vision
- —
- Streaming
- ✓
- License
- proprietary
- Released
- 2025-01-31
Pricing
- Input / 1M
- $1.10
- Output / 1M
- $4.40
- Cached input / 1M
- $0.55
Cost estimate
o3-mini is OpenAI's small reasoning model in the o-series, optimised for math, science, and coding with adjustable reasoning effort (low/medium/high). It supports function calling and structured outputs but is text-only — no vision, no image input. At $1.1/$4.4 per million tokens with a 200K context, it's positioned as a cheap default for agent loops and code tasks where you want chain-of-thought without paying o1 prices.
Editor's verdict
Pick o3-mini when you need o-series reasoning quality on a budget — it beats GPT-4o-class models on math and code benchmarks at a fraction of o1's cost. The catch: no vision, and on open-ended writing or world-knowledge tasks it's noticeably weaker than GPT-4o or Claude Sonnet. For multimodal or general chat, look elsewhere; for coding agents and structured problem-solving, it's a solid default until o3 or o4-mini land.
Reviews
No reviews yet. Be the first.
Last updated: 2026-04-29