OpenAI model product
o3
OpenAI's most powerful reasoning model.
Updated Aug 12, 2026. Default version: o3
LLMBoard score52.6o3
Coverage100%22 benchmark families
Context window200KTokens
Official input price$2OpenAI API
o3 Specifications
Technical details for the model's default version.
- Version
- o3
- Released
- Apr 16, 2025
- Knowledge cutoff
- May 31, 2024
- Parameters
- N/A
- Context window
- 200K
- Max output
- 100K
- Inputs
- image, pdf, text
- Outputs
- text
- Open weights
- No
- License
- Proprietary
o3 Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
o3 Benchmark Results
Benchmark scores for o3.
| Aider-Polyglot | 81.3% | 03 | 22 | 90.5% | C | |
| COLLIE | 98.4% | 03 | 10 | 77.8% | C | |
| MathVista | 86.8% | 03 | 39 | 94.7% | C | |
| ARC-AGI | 88.0% | 04 | 7 | 50.0% | C | |
| AIME 2024 | 91.6% | 06 | 53 | 90.4% | C | |
| Tau-bench | 63.0% | 06 | 6 | 0.0% | C | |
| MMMU | 82.9% | 07 | 63 | 90.3% | C | |
| Tau2 Retail | 80.2% | 08 | 26 | 72.0% | C | |
| ERQA | 64.0% | 09 | 23 | 63.6% | C | |
| Tau2 Airline | 64.8% | 09 | 23 | 63.6% | C | |
| Multi-Challenge | 60.4% | 10 | 29 | 67.9% | C | |
| FrontierMath | 15.8% | 13 | 17 | 25.0% | C | |
| VideoMMMU | 83.3% | 13 | 26 | 52.0% | C | |
| ARC-AGI v2 | 6.5% | 15 | 16 | 6.7% | B | |
| CharXiv-R | 78.6% | 23 | 48 | 53.2% | C | |
| MMMU-Pro | 76.4% | 28 | 66 | 58.5% | C | |
| Tau2 Telecom | 58.2% | 30 | 35 | 14.7% | C | |
| BrowseComp | 49.7% | 44 | 58 | 24.6% | C | |
| AIME 2025 | 86.4% | 58 | 114 | 49.6% | C | |
| SWE-Bench Verified | 69.1% | 65 | 105 | 38.5% | C | |
| GPQA | 83.3% | 67 | 234 | 71.7% | C | |
| Humanity's Last Exam | 14.7% | 72 | 93 | 22.8% | C |
o3 Arena Results
Preference and agent-evaluation results for the default version.
| search style control | overall | 15 | 1190.9 | 20,785 | N/A | |
| search | overall | 24 | 1143.5 | 20,785 | N/A | |
| search factuality | overall | 30 | 1128.1 | 3,706 | N/A | |
| vision style control | overall | 62 | 1216.8 | 46,095 | N/A | |
| vision | overall | 65 | 1215.3 | 46,095 | N/A | |
| text style control | overall | 91 | 1430.6 | 59,902 | N/A | |
| text | overall | 119 | 1409.1 | 59,902 | N/A | |
| text factuality | overall | 129 | 1403.2 | 7,585 | N/A |
o3 Runtime Performance
Provider-specific output speed and catalog latency for o3. Runtime does not affect the capability score.
| OpenAI | 50 tok/s | 20 s | 200K | 100K |
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
o3 Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $2 input, $8 output per 1M
- Official provider
- OpenAI
- Lowest third-party
- From $1.8 input, $7.2 output per 1M via Poe
- Tracked offerings
- 11
| Poe | openai/o3 | global | $1.8 | $7.2 | 200K | |
| Impossibl | openai/o3 | global | $2 | $8 | 200K | |
| NanoGPT | openai/o3 | global | $2 | $8 | 200K | |
| NEAR AI Cloud | openai/o3 | global | $2 | $8 | 200K | |
| Kilo Gateway | openai/o3 | global | $2 | $8 | 200K | |
| Merge Gateway | openai/o3 | global | $2 | $8 | 200K | |
| Vercel AI Gateway | openai/o3 | global | $2 | $8 | 200K | |
| OpenRouter | openai/o3 | global | $2 | $8 | 200K | |
| Cloudflare AI Gateway | openai/o3 | global | $2 | $8 | 200K | |
| OpenAI | o3 | global | $2 | $8 | 200K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
o3 Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
| o3 | 52.6 | N/A | 200K | 100K | No | Proprietary |
What is o3?
Key information about o3 and its available data.
OpenAI's most powerful reasoning model. o3 is a well-rounded and powerful model across domains. It sets a new standard for math, science, coding, and visual reasoning tasks. It also excels at technical writing and instruction-following.
Use it to think through multi-step problems that involve analysis across text, code, and images.
Data as of 2026-08-11.
o3 vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to o3
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
OP
GPT-5.5-Instant
OpenAI
OP
GPT-5
OpenAI
OP
GPT-5.1-Codex
OpenAI
OP
GPT-5.4-nano
OpenAI
OP
GPT-OSS-20B
OpenAI
OP
GPT-5.2-Codex
OpenAI
FAQ
Common questions about o3.
When was o3 released?
o3's default version was released on Apr 16, 2025.
How much does o3 cost?
o3's official API price is $2 per million input tokens and $8 per million output tokens via OpenAI. The lowest tracked third-party offer starts at $1.8 input and $7.2 output via Poe.
Who created o3?
o3 was created by OpenAI.
What is the context window for o3?
The default version has a 200K token context window.
Is o3 open weight?
No. The default version is not marked as having publicly available weights.
How many API providers offer o3?
11 provider offerings are linked to the default version.
What models should I compare o3 with?
Nearby ranked alternatives include Seed 2.0 Lite, MiMo V2 Omni, Gemma 4 26B A4B.