o3
Reasoning model by OpenAI.
April 16, 2025 checkpoint of OpenAI o3, setting high benchmarks on mathematics, code generation, and visual reasoning across a 200,000-token context window with up to 100,000 generation tokens.
This model is deprecated
o3 has been deprecated. It is still served on nRouter and can still be called.
Caller Notice: To help identify integrations still relying on this model, all API responses now include x-nr-model-deprecated: true and Warning: 299 - "Model deprecated" headers.
Specifications
Context window, ranked against other reasoning models
Proportional scale: 0 to 200K tokens. Not ranked, no published context window: gpt-5.2-pro-2025-12-11, gpt-5.4-pro, grok-4-1-fast-non-reasoning, grok-4-1-fast-reasoning, gpt-5.2-pro.
—
—
reasoning
Benchmarks
Pricing
Rates are read live
Pricing for o3-2025-04-16 is fetched from the live catalogue on load, so it is never served from a cache that could outlive a repricing.
Availability
Not enough data yet
We have not collected enough health probes for this model to publish an uptime or latency figure. A percentage from a handful of samples is fabricated precision, so none is shown until the sample count supports one.
How o3 is billed
Billed at the provider's list price, with no per-token markup.
- Exact list price
- Tokens settle at the published rate of the model that served your call.
- Cache discounts passed through
- Provider prompt-caching discounts reach your balance with no added fee.
- Cost on every response
- Each response carries the
x-nr-request-costheader, so you can reconcile spend per request.
More from OpenAI
Start building with o3-2025-04-16
One API key, three steps. Choose your tool to begin.
Get a key
Create API keySet your API key
export NROUTER_API_KEY="<YOUR_API_KEY>"Configure Claude Code
Choose a tool, then copy its setup from the panel.
FAQs
What is o3?
April 16, 2025 checkpoint of OpenAI o3, setting high benchmarks on mathematics, code generation, and visual reasoning across a 200,000-token context window with up to 100,000 generation tokens.
How much does o3 cost?
o3 is billed at the provider's list price on nRouter with no per-token markup — tokens settle at the exact upstream rate. Input and output rates are read live, so see the pricing section above for the current figures.
Does o3 support tool calling and structured outputs?
Yes. o3 supports tool calling (function calling), so it can invoke external tools and APIs during a conversation.
What inputs and outputs does o3 support?
o3 takes text and images as input and returns text as output.
How do I use o3 via nRouter?
Use any OpenAI-compatible client: point it at https://api.nrouter.ai/v1 with your nRouter API key and set the model to "openai/o3-2025-04-16". The quick-start panel above has a copy-paste curl command for this model.
When was o3 released?
o3 was released on April 16, 2025.
Where is the o3 documentation?
Start at nrouter.ai/docs. The quick-start guide (nrouter.ai/docs/getting-started/quick-start) covers API keys and your first request, and the API reference (nrouter.ai/docs/api-reference) lists every parameter.
What other text models does OpenAI have?
Other OpenAI text models on nRouter include GPT-OSS 120B 1 (nrouter.ai/models/gpt-oss-120b-1), GPT-6.1 Sol (nrouter.ai/models/gpt-6.1-sol), GPT-5.2 Pro (nrouter.ai/models/gpt-5.2-pro-2025-12-11), GPT Audio 1.5 (nrouter.ai/models/gpt-audio-1.5).