Tool support requires the appropriate API integration; a supported tool is not automatically active in every chat.
Supported API features and tools
Function calling and structured outputs
Text and image input
Web search, file search, image generation tools, and MCP
Batch requests through a supported API integration
Before you choose
No native audio or video support.
Streaming is not supported; plan for a completed response rather than token-by-token output.
Code interpreter and fine-tuning are not supported.
Requests can take several minutes. The o3-pro-2025-06-10 snapshot is scheduled to retire December 11, 2026.
Reasoning behavior
o3-pro uses additional compute for extended reasoning, but its reviewed model reference does not enumerate adjustable effort levels or a default. Do not assume it shares o3’s controls or GPT-5 Pro’s high-only setting. Check the current API contract before configuring an integration.
Responses API only; Chat Completions is not supported. Batch is listed as supported.
Consider background mode for long-running requests. It is an asynchronous API workflow, not streaming output.
The documented snapshot is o3-pro-2025-06-10. API rate limits depend on usage tier; the free API tier is not supported.
Put it to work
Start with a more useful prompt.
Original examples from EZ Ai Assist. Adapt these to your task and the features available in your workspace.
Workflow 01
Stress-test an engineering proposal
Focus a slow, careful review on assumptions and failure conditions.
Stress-test this engineering proposal against its stated requirements and constraints. For each major design choice, identify the strongest counterargument supported by the supplied evidence and a scenario in which it would fail. Separate confirmed contradictions from unresolved risks. Recommend the next validation step and acceptance criteria; do not redesign components outside the proposal.
Workflow 02
Reconcile conflicting analyses
Compare evidence and assumptions before selecting a conclusion.
Compare these two analyses of the same problem. Map where they agree, where their data or assumptions differ, and which conclusions cannot both hold. Cite the relevant passages or calculations. Explain what additional evidence would resolve each disagreement and produce a conditional recommendation rather than forcing certainty where the source material is incomplete.
Workflow 03
Audit a technical proof
Ask for checkable objections instead of a confident verdict alone.
Audit the proposed proof below using the stated definitions and domain assumptions. Check each implication, boundary case, and use of a theorem. For a suspected gap, explain why it matters and provide a counterexample if the supplied assumptions permit one. Distinguish a proven error from an unverified step, then propose the smallest correction that would make the argument checkable.
Developer reference
OpenAI API pricing
These are OpenAI API reference prices, not EZ Ai Assist subscription prices.
These are Standard rates for o3-pro, not o3. The reviewed table lists no cached-input discount or separate long-context price.
Additional reasoning contributes to output usage even when the visible answer is short.
Tools and other service tiers have separate pricing. Verify actual usage and current rates when evaluating a replacement.
Common questions
A few things worth knowing.
What is retiring, and when?
The announced retirement concerns o3-pro-2025-06-10, with shutdown scheduled for December 11, 2026 following the June 11 announcement. Check the linked OpenAI notice and test a replacement before relying on continued API access.
What replaces o3-pro?
The deprecation notice recommends GPT-5.6 Sol with reasoning.mode: pro in the API. This mode is not an EZ Ai Assist Pro subscription, and app controls may differ. Compare outputs and integration behavior before migrating.
Can o3-pro stream its answer?
No. The model reference marks streaming unsupported. Background mode can help an API integration handle a long-running request, but it does not turn this model into a token-streaming model.
Can I choose low, medium, or high reasoning?
The reviewed o3-pro reference does not establish a list of effort choices or a default. It describes extended reasoning, which is not evidence for inheriting another model’s parameter settings. Check the current API contract rather than guessing.
Does it support structured outputs and tools?
Yes, structured outputs and function calling are listed, alongside web search, file search, image generation tools, and MCP. Code interpreter is not supported. Tool availability in a provider API is separate from what the app exposes.
Why can a short answer still be expensive?
The model uses reasoning in addition to visible output. Standard reference rates are $20 input and $80 output per million tokens, with no cached-input discount listed. Measure billed usage and tool charges rather than estimating solely from the final paragraph.
Can I use it for time-sensitive chat?
Some requests can take minutes, so evaluate latency before relying on it for interactive work. The prompts here suit reviewable analysis, not guaranteed instant answers. Check current EZ Ai Assist availability and verify any important conclusion.
Check the source
Official documentation
Specifications and API prices checked on . Example prompts and workflow advice are editorial guidance from EZ Ai Assist.