Home/Models/o3

o3

by OpenAI· Released 2025-04
Official Docs ↗

o3 is OpenAI's mainstream reasoning model, replacing the original o1 with a dramatic price cut and better benchmark scores. It is the recommended choice for math, science, and complex agentic tasks where extended thinking improves quality.

Input Price
$2.00/1M
per 1M tokens
Output Price
$8.00/1M
per 1M tokens
Context Window
200K tokens
max tokens
ELO Score
1395
LMSYS Arena

Pricing Breakdown

VolumeInput CostOutput CostCombined (50/50)
1,000 tokens$0.0020$0.0080$0.0050
10,000 tokens$0.0200$0.0800$0.0500
100,000 tokens$0.2000$0.8000$0.5000
1,000,000 tokens$2.0000$8.0000$5.0000

Strengths

  • Top-tier reasoning on math, science, and agentic tasks
  • 87% cheaper than original o1 at the same capability tier
  • 200K context window
  • Strong chain-of-thought for complex problem solving

Weaknesses

  • Higher latency than non-reasoning models
  • Verbose output increases effective cost

Best For

reasoningmathresearchcoding

Compare o3 With

o3 vs GPT-5.4
OpenAI
o3 vs Claude Opus 4.7
Anthropic
o3 vs Gemini 3.1 Pro
Google
o3 vs Claude Sonnet 4.6
Anthropic
o3 vs Gemini 2.5 Pro
Google
o3 vs DeepSeek V3.2 (Reasoner)
DeepSeek

Frequently Asked Questions

How much does o3 cost?

o3 costs $2.00/1M per 1M input tokens and $8.00/1M per 1M output tokens.

What is o3's context window?

o3 has a context window of 200K tokens, which means it can process up to 200,000 tokens in a single request.

Is o3 open source?

No, o3 is a proprietary model by OpenAI and is not open source.

What is o3 best used for?

o3 is best suited for: reasoning, math, research, coding. Top-tier reasoning on math, science, and agentic tasks.

What is o3's ELO score?

o3 has an ELO score of 1395 on the LMSYS Chatbot Arena leaderboard, placing it among top-tier models.