o1
OpenAI o1: the first o-series model trained with RL for internal chain-of-thought reasoning. Features a 200K context window and image input support.
Model Overview
Capabilities, design details, and architectural traits
o1 - The First Full Public o-Series Model with an Internal COT
o1 is the foundational full-scale model in OpenAI's o-series. It is trained with reinforcement learning to produce a long internal chain of thought before generating a response. This pre-answer reasoning step is the defining mechanism of the o-series line and was first delivered at full scale through o1.
| Trait | Detail |
|---|---|
| Internal chain of thought | Models think before answering, generating a private reasoning sequence; this reasoning is tracked as reasoning tokens separate from visible output tokens |
| Training method | Trained with reinforcement learning to perform complex reasoning - not standard supervised fine-tuning |
| Fine-tuning | Not supported |
| Predicted outputs | Not supported |
How o1 Differs from the o-Series That Followed
o1 established the o-series pattern: RL-trained chain-of-thought reasoning before each answer, slowest speed tier, no fine-tuning support, and no predicted outputs. Later models in the series (o3, o3-pro) carry forward this pattern with additional capabilities such as image integration into the reasoning chain and RL-trained tool use. o1 does not include those extensions.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | OpenAI | Anthropic | Anthropic |
| Release Date | September 12, 2024 | September 22, 2026 | September 28, 2026 |
| Knowledge Cutoff | Oct 2023 | - | Jun 2026 |
| Context & Limits | |||
| Context Window | 200K | 1M Best Context Window | 1M Best Context Window |
| Pricing (per 1M tokens) | |||
| Input Pricing | $15 | $4 | $2 Best Input Pricing |
| Output Pricing | $60 | $20 | $10 Best Output Pricing |
| Modalities | |||
| Inputs | textimagefile | textimagefile | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 15.2 | 57.6 Best Intelligence Index | 56.0 |
| Coding Index | 39.7 | - | - |
| Agentic Index | - | - | - |
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
SciCode Benchmark
Scientific coding and mathematical modeling.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
Explore more from OpenAI
Other models by OpenAI
Top AI Models
Leading alternatives by intelligence score
