o3-pro is an extended-thinking version built by OpenAI on o3, focused not on rushing to provide answers, but on dedicating more reasoning to complex problems. It supports text and image understanding, making it suitable for mathematical proofs, logical analysis, and multi-step solution arguments. Applications can integrate it using the public request format in this page's API section.
Point the Base URL to api.acedata.cloud, configure your platform API key and the model ID below, and use your compatible SDK or client.
import os
from openai import OpenAI
client = OpenAI(
api_key=os.environ["ACEDATACLOUD_API_KEY"],
base_url="https://api.acedata.cloud/v1",
)
response = client.responses.create(
model="o3-pro",
input="Hello!",
)
print(response.output_text)Choose an available protocol for this model. OpenAI SDK uses a Base URL ending in /v1; Anthropic SDK uses the root URL. See each guide for protocol-specific parameters, tools and response formats.
Clarify capacity, input and output, and calling methods before selecting a model.
Extended thinking is o3-pro's positioning. Applications should organize relevant inputs and history according to the public protocol, and verify native Pro specifications and platform-compatible formats separately; the presence of a field does not mean all values apply.
Learn what o3-pro can bring to your work.
The core value of o3-pro lies in extended thinking, making it suitable for handling intertwined conditions and problems that cannot be solved by directly looking up an answer. You can ask it to distinguish premises, derivations, and conclusions, and list counterexamples or failure conditions, making the deliverable an analysis that can be reviewed item by item rather than merely an apparently certain judgment.
Text and images can jointly form task input, for example, using a diagram to supplement positional relationships and then using text to explain the conditions that need verification. o3-pro can generate explanations and reasoning results based on information in the image. It is suitable for turning charts, sketches, or screenshots of questions into analytical material, rather than treating visual capability as a drawing feature.
Mathematical proofs, scientific arguments, and complex plans are best advanced in stages: first confirm definitions and constraints, then examine intermediate propositions, and finally summarize answers and unproven items. Responses is suitable for organizing this kind of input and results; each addition should explain which premises have changed to avoid continuing to rely on invalid arguments.
Start with specific tasks to find where the model can be effective.
Enter the problem, known conditions, and an existing proof, and ask o3-pro to check implicit assumptions, whether case discussions are complete, and whether the conclusion truly follows from the premises. Deliverables may include a proof draft, a list of concerns, and steps to be verified; if the problem includes a geometric diagram, an illustration can also be attached to help explain relationships between objects.
Submit the plan's objectives, resource constraints, and candidate paths, have the model reason through outcomes under different assumptions, and then organize the trade-off rationale and key dependencies. This is suitable for preparing an argument before engineering design or business decisions, especially for tasks where multiple conditions affect one another and it is necessary to explain why a particular path was chosen.
Enter charts or flowcharts together with explanatory text, and ask o3-pro to cross-check whether the textual conclusions align with the information in the visuals, while identifying ambiguities, omissions, or data that needs to be added. The final output can be an itemized verification report specifying which conclusions are supported by the available materials and which still require additional validation.
Choose based on task complexity, input materials, and expected results.
o3-pro is not another way of writing o3; it is a version designed for longer reasoning. When a task involves complex constraints, proof review, or verification of important conclusions, and sufficient time for analysis is available, o3-pro can be considered first; for general analysis, you can start with o3 to establish a baseline, then compare results using the same set of real problems, without needing to upgrade every task.
o4-mini is positioned more toward fast, cost-efficient reasoning, making it suitable as a candidate for a large number of everyday tasks; o3-pro is better suited to complex problems where analytical quality is the top priority. You can use a smaller model for preliminary organization, then hand disputed points to o3-pro for in-depth argumentation. The choice should be based on task difficulty and acceptable wait time, rather than model names alone.
Based on the characteristics of o3-pro, first validate small tasks whose results can be checked.
You can ask directly: Review this proof, check step by step whether the premises are sufficient, identify leaps in reasoning, and provide supplementary arguments or counterexamples. Leave propositions that cannot be confirmed as items to be proved.
Provide complete definitions and prerequisite propositions; allow waiting time for longer reasoning, and independently verify key conclusions.
Use the full model ID o3-pro, first confirm the public request format and available parameters on the API page, then connect your application. Retain result parsing, exception handling, and relevant evidence, and use the same set of real samples to assess whether it is suitable for continued use.
Before formal use, understand the quality of outputs and the scope of capabilities.
Answers to common questions about using o3-pro.
No. o3-pro is OpenAI's publicly available extended-thinking version of o3, designed to provide more reliable answers; use o3-pro when calling it. It is better suited to complex reasoning and verification, but that does not mean every simple question needs it, nor that all tasks will see the same degree of improvement.
Submit clear images and text questions according to the image-and-text format of the selected public API. Chat Completions uses text and image_url; Responses uses the corresponding image input content blocks. Clearly specify the areas of focus and expected output, and verify key numbers and image details against the original image.
Use Chat Completions or Responses and provide the complete model ID. Chat Completions uses messages and choices, while Responses uses input and the corresponding response structure; handle history management, streaming events, and tool parameters separately according to the selected API, and do not mix the two formats.
When using Chat Completions, place the relevant history in messages; when using Responses, organize input and related conversation content according to the documentation. Provide the latest materials, revision goals, and key constraints in each turn; for longer tasks, retain interim summaries and a final version that can be checked independently.
Output length control and the model's extended-thinking purpose are not the same thing. max_output_tokens is used to constrain the response budget and should not be regarded as a quality guarantee. A more effective approach is to provide complete conditions, clearly state verification goals, and request that assumptions be distinguished from conclusions; reasoning controls cannot simply copy values from other models either.
Choose a development approach or application suited to this model and follow its setup guide.
↗OpenAI Chat Completions API
SDK installation, authentication and examples↗OpenAI Responses API
SDK installation, authentication and examples
↗Using Ace Data Cloud with Cherry Studio
Chat client setup · Base URL, API key and model
↗Using Ace Data Cloud with ChatBox
Chat client setup · Base URL, API key and modelThis is only a basic call example. See the full docs for more parameters and advanced usage.
View Docs