o3: Pricing, Context, and Benchmarks

o3 is OpenAI's o-series model released on 2025-04-16. It runs on a 200K-token context window, priced at $2.00/M input and $8.00/M output. Reasoning-trained model. Higher per-call cost than GPT-4o because internal reasoning tokens are billed.

Specifications

Benchmark scores

Scores reflect single measurements on the date shown. Reproducibility varies; consult the linked source for methodology. Use the comparison tool for side-by-side analysis.

What this model is good for

Based on benchmark scores and capability fit, o3 ranks well for:

• Math & STEM Reasoning — rank #3. Reasoning-trained o-series. AIME 2025: 88.9%.

Cost calculator

For 1,000 tokens of input and 1,000 tokens of output per call, this model costs $0.0100 per call. For typical usage:

For more precise estimates including caching discounts, use the LLM API Cost Calculator.

Other OpenAI models

• GPT-5
• GPT-5 Mini
• GPT-5 Nano
• GPT-4.1
• o3-mini
• GPT-4o Mini

Common comparisons

• o3 vs Mistral Large 2 Side-by-side specs, pricing, and benchmarks
• o3 vs Qwen3-Coder-480B Side-by-side specs, pricing, and benchmarks
• o3 vs Command R+ Side-by-side specs, pricing, and benchmarks
• o3 vs Gemini 2.5 Pro Side-by-side specs, pricing, and benchmarks

API setup

curl https://api.openai.com/v1/responses \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "o3", "messages": [{"role": "user", "content": "Hello"}]}'

See the official documentation for the full request/response schema. For pre-flight token counting before paying, use the appropriate token counter.