o3
OpenAI · Released 16 Apr 2025 · o3
$2.00
Input price / 1M tokens
#36 of 61, cheapest first
$8.00
Output price / 1M tokens
#37 of 61, cheapest first
200K
Context window
#62 of 71
102
Output tokens / second
#29 of 61
69.1%
SWE-bench Verified
#8 of 12
98.4%
AIME 2025
#3 of 7
Summary
o3 is a proprietary model from OpenAI, released on 16 Apr 2025.
At $2.00 input and $8.00 output per million tokens, it is #36 of 61 on input price, cheapest first.
Its 200K-token context window ranks #62 of 71.
Measured output speed is 102 tokens per second, #29 of 61.
An o-series reasoning model for math, science, coding and visual reasoning. The docs say GPT-5 succeeded it.
Benchmarks · 2 reported
| Benchmark | Score | Bar | Rank | Source |
|---|---|---|---|---|
| SWE-bench Verifiedn=477 subset; high reasoning effort | 69.1% | #8 of 12 | OpenAI GPT-5 for developers post (comparison)Self-reported | |
| AIME 2025pass@1 with Python interpreter | 98.4% | #3 of 7 | OpenAI launch postSelf-reported |
Self-reported means the lab ran the test itself. A rank appears only where several labs report the same version of a benchmark.
Similar models
Models with the most benchmarks in common with o3, and the closest scores.
SWE-bench Verified
#8 of 12Real GitHub issues, fixed and tested
- o369.1
- gpt-oss-120b62.4
- gpt-oss-20b60.7
- Gemini 3 Flash78
- Gemini 2.5 Pro59.6
- Gemini 2.5 Flash60.4
AIME 2025
#3 of 7Competition mathematics
- o398.4
- gpt-oss-120b97.9
- gpt-oss-20b98.7
- Gemini 3 Flash95.2
- Gemini 2.5 Pro88
- Gemini 2.5 Flash72
Compare o3 side by side
Pre-filled with the two closest models. Swap any of them.
Sources
- price, tiers: developers.openai.com/api/docs/pricing, checked 2 Oct 2026
- context, max output, knowledge cutoff (Jun 1 2024), snapshot: developers.openai.com/api/docs/models/o3, checked 2 Oct 2026
- shutdown date Dec 11 2026, replacement: developers.openai.com/api/docs/deprecations, checked 2 Oct 2026
- release date Apr 16 2025, AIME result: openai.com/index/introducing-o3-and-o4-mini/, checked 2 Oct 2026
- output speed: artificialanalysis.ai/models/o3, checked 2 Oct 2026
Building on o3?
Get the architecture right before the bill arrives.
We size caching, routing and fallbacks for your workload on a free call, and tell you where a cheaper model is good enough.