OpenAI model product
o1
A research preview model focused on mathematical and logical reasoning capabilities, demonstrating improved performance on tasks requiring step-by-step reasoning, mathematical problem-solving, and code generation.
Updated Aug 12, 2026. Default version: o1
LLMBoard score36.3o1
Coverage80%16 benchmark families
Context window200KTokens
Official input price$15OpenAI API
o1 Specifications
Technical details for the model's default version.
- Version
- o1
- Released
- Dec 17, 2024
- Knowledge cutoff
- Sep 1, 2023
- Parameters
- N/A
- Context window
- 200K
- Max output
- 100K
- Inputs
- image, pdf, text
- Outputs
- text
- Open weights
- No
- License
- Proprietary
o1 Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
o1 Benchmark Results
Benchmark scores for o1.
| GPQA Biology | 69.2% | 01 | 1 | 100.0% | C | |
| GPQA Chemistry | 64.7% | 01 | 1 | 100.0% | C | |
| GPQA Physics | 92.8% | 01 | 1 | 100.0% | C | |
| MATH | 96.4% | 02 | 71 | 98.6% | C | |
| MMLU | 91.8% | 02 | 100 | 99.0% | C | |
| GSM8k | 97.1% | 03 | 48 | 95.7% | C | |
| MGSM | 89.3% | 10 | 31 | 70.0% | C | |
| TAU-bench Retail | 70.8% | 10 | 25 | 62.5% | C | |
| MathVista | 71.8% | 12 | 39 | 71.0% | C | |
| TAU-bench Airline | 50.0% | 12 | 23 | 50.0% | C | |
| FrontierMath | 5.5% | 17 | 17 | 0.0% | C | |
| SimpleQA | 47.0% | 17 | 46 | 64.4% | C | |
| MMMU | 77.6% | 18 | 63 | 72.6% | C | |
| MMMLU | 87.7% | 19 | 49 | 62.5% | C | |
| HumanEval | 88.1% | 26 | 66 | 61.5% | C | |
| LiveBench | 67.0% | 33 | 38 | 13.5% | C | |
| AIME 2024 | 74.3% | 35 | 53 | 34.6% | C | |
| SWE-Bench Verified | 41.0% | 95 | 105 | 9.6% | C | |
| GPQA | 78.0% | 96 | 234 | 59.2% | C |
o1 Arena Results
Preference and agent-evaluation results for the default version.
| vision style control | overall | 76 | 1193.4 | 3,694 | N/A | |
| vision | overall | 89 | 1168.6 | 3,694 | N/A | |
| text style control | overall | 135 | 1402.2 | 27,807 | N/A | |
| text | overall | 164 | 1365.9 | 27,807 | N/A |
o1 Runtime Performance
Provider-specific output speed and catalog latency for o1. Runtime does not affect the capability score.
| OpenAI | 66 tok/s | 16.2 s | 200K | 100K | |
| Azure | 16 tok/s | 0.54 s | 200K | 100K |
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
o1 Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $15 input, $60 output per 1M
- Official provider
- OpenAI
- Lowest third-party
- N/A
- Tracked offerings
- 1
| OpenAI | o1 | global | $15 | $60 | 200K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
o1 Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
| o1 | 36.3 | N/A | 200K | 100K | No | Proprietary | |
| o1-preview | N/A | N/A | 128K | 32.8K | No | Proprietary |
What is o1?
Key information about o1 and its available data.
A research preview model focused on mathematical and logical reasoning capabilities, demonstrating improved performance on tasks requiring step-by-step reasoning, mathematical problem-solving, and code generation. The model shows enhanced capabilities in formal reasoning while maintaining strong general capabilities.
Data as of 2026-08-11.
o1 vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to o1
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
OP
o3 mini
OpenAI
OP
GPT-4.5
OpenAI
OP
o1 pro
OpenAI
OP
GPT-4.1
OpenAI
OP
GPT-5-nano
OpenAI
OP
GPT-5.3
OpenAI
FAQ
Common questions about o1.
When was o1 released?
o1's default version was released on Dec 17, 2024.
How much does o1 cost?
o1's official API price is $15 per million input tokens and $60 per million output tokens via OpenAI.
Who created o1?
o1 was created by OpenAI.
What is the context window for o1?
The default version has a 200K token context window.
Is o1 open weight?
No. The default version is not marked as having publicly available weights.
How many API providers offer o1?
1 provider offerings are linked to the default version.
What models should I compare o1 with?
Nearby ranked alternatives include MiniMax M1 80K, LongCat Flash Chat, Qwen3 Next 80B A3B.