OpenAI: o3
The OpenAI o3 model is a language processing tool that can handle various input types, including text, images, and files, with a context length of up to 200,000 tokens. It supports tools and reasoning capabilities, but does not produce structured output. This model may be worth considering for those who require advanced reasoning abilities or need to integrate with external tools, given its moderate price point of $2.0 per million input tokens and $8.0 per million output tokens. Its blended benchmark score is 72.1 across three independent benchmarks, which may indicate a balance between performance and generalizability. However, further evaluation would be necessary to fully assess the model's suitability for specific use cases.
Benchmark results
Independent, published benchmarks. Blended score 71.2 across 3 benchmarks, last refreshed 2026-09-27. How scoring works →
| Benchmark | Measures | Score |
|---|---|---|
| Aider Polyglot | code editing across languages | 81.3 |
| AI Index | broad capability composite | 33.3 |
| EQ-Bench | emotional understanding in dialogue | 74.7 |
- Model ID
- openai/o3
- Vendor
- openai
- Released
- April 2025
- Tokenizer
- GPT
- Input Modalities
- image, text, file
- Output Modalities
- text
- Max Output
- 100,000 tokens
- Tool Calling
- ✓ supported
- Structured Output
- ✓ supported
- Reasoning Mode
- ✓ supported
- Vision
- ✓ accepts images
- Audio
- no
- Moderated
- yes
What it costs in practice
Computed from the current $2.00/M input and $8.00/M output rates. Run your own numbers →
| Job | Tokens | Cost |
|---|---|---|
| Summarize a 50-page report | 30k in / 1.5k out | $0.07 |
| Classify 1,000 customer emails | 500k in / 50k out | $1.40 |
| A month of a busy support chatbot | 5M in / 2M out | $26.00 |
Strong choice for
Regex Writing
Frontend Component Design
Data Analysis
Marketing Copy
Language Translation
Financial Analysis
Category rankings
Where OpenAI: o3 places across the 55 categories it ranks in. How we rank →
| # | Category | Score |
|---|---|---|
| #2 | Data AnalysisData · of 25 ranked | 148 |
| #2 | Financial AnalysisProfessional · of 25 ranked | 143 |
| #2 | Math TutoringEducation · of 25 ranked | 137 |
| #2 | Diagram ExtractionVision · of 25 ranked | 139 |
| #2 | Browser AutomationAgents · of 25 ranked | 148 |
| #2 | Fitness CoachingPersonal · of 25 ranked | 127 |
| #2 | Experiment DesignResearch · of 25 ranked | 144 |
| #3 | SOP / Process DocsBusiness · of 25 ranked | 128 |
| #4 | Regex WritingCode · of 25 ranked | 130 |
| #4 | Frontend Component DesignCode · of 25 ranked | 137 |
| #4 | Marketing CopyWriting · of 25 ranked | 125 |
| #4 | Language TranslationWriting · of 25 ranked | 130 |
| #4 | Standardized Test PrepEducation · of 25 ranked | 125 |
| #4 | Resume WritingBusiness · of 25 ranked | 115 |
| #4 | Screenshot DebuggingVision · of 25 ranked | 136 |
| #4 | Chart & Graph ReadingVision · of 25 ranked | 143 |
| #4 | Tutoring for KidsPersonal · of 25 ranked | 130 |
| #5 | Sales / Cold EmailBusiness · of 25 ranked | 120 |
| #5 | Job Application DraftingBusiness · of 25 ranked | 120 |
| #5 | Function / Tool CallingAgents · of 25 ranked | 129 |
| #5 | Journaling HelperPersonal · of 25 ranked | 120 |
| #5 | Recipe GenerationPersonal · of 25 ranked | 120 |
| #6 | Physics TutoringEducation · of 25 ranked | 139 |
| #7 | Unit Test GenerationCode · of 25 ranked | 141 |
| #7 | CI/CD PipelinesCode · of 25 ranked | 135 |
| #7 | Agent WorkflowsAgents · of 25 ranked | 163 |
| #7 | Coding AgentsAgents · of 25 ranked | 163 |
| #8 | SQL GenerationCode · of 25 ranked | 152 |
| #11 | Code ReviewCode · of 25 ranked | 153 |
| #11 | Bug FixingCode · of 25 ranked | 157 |
| #11 | Essay GradingEducation · of 25 ranked | 153 |
| #11 | Scientific CodingResearch · of 25 ranked | 157 |
| #12 | Blog Post WritingWriting · of 25 ranked | 133 |
| #12 | Trip PlanningPersonal · of 25 ranked | 133 |
| #15 | Creative WritingWriting · of 25 ranked | 146 |
| #15 | Screenwriting & DialogWriting · of 25 ranked | 146 |
| #15 | Academic WritingWriting · of 25 ranked | 146 |
| #15 | Legal DraftingProfessional · of 25 ranked | 146 |
| #15 | OKRs & Strategic PlanningBusiness · of 25 ranked | 146 |
| #15 | Math ProofsResearch · of 25 ranked | 146 |
| #16 | ETL ScriptingData · of 25 ranked | 142 |
| #16 | Medical Note SummarizationProfessional · of 25 ranked | 142 |
| #17 | OCR / Document ParsingData · of 25 ranked | 136 |
| #17 | Table Extraction from PDFsData · of 25 ranked | 136 |
| #21 | CSV / Spreadsheet CleanupData · of 25 ranked | 139 |
| #21 | RFP ResponseBusiness · of 25 ranked | 156 |
| #22 | Contract ReviewProfessional · of 25 ranked | 160 |
| #23 | Legal ResearchProfessional · of 25 ranked | 154 |
| #23 | Scientific ResearchProfessional · of 25 ranked | 154 |
| #23 | History TutoringEducation · of 25 ranked | 136 |
| #23 | Character RoleplayPersonal · of 25 ranked | 141 |
| #23 | Fiction CollaboratorPersonal · of 25 ranked | 154 |
| #23 | Literature ReviewResearch · of 25 ranked | 154 |
| #25 | Code RefactoringCode · of 25 ranked | 150 |
| #25 | RAG PipelinesAgents · of 25 ranked | 150 |
Similar models
OpenAI: GPT-6 Luna Pro
OpenAI: GPT-6 Luna Pro (batch)
OpenAI: GPT-6 Luna
OpenAI: GPT-6 Luna (batch)
OpenAI: GPT-6 Sol Pro
OpenAI: GPT-6 Sol Pro (batch)
Quick answers
- How much does OpenAI: o3 cost?
- $2.00 per million input tokens and $8.00 per million output tokens.
- What is OpenAI: o3's context window?
- 200,000 tokens, roughly 300 pages of text in a single request.
- Does OpenAI: o3 support tool calling?
- It supports tool calling, structured output, a reasoning mode.
- Can OpenAI: o3 process images?
- Yes, it accepts image input.