gpt-5.5
GPT‑5.5 understands what you’re trying to do faster and can carry more of the work itself. It excels at writing and debugging code, researching online, analyzing data, creating documents and spreadsheets, operating software, and moving across tools until a task is finished. Instead of carefully managing every step, you can give GPT‑5.5 a messy, multi-part task and trust it to plan, use tools, check its work, navigate through ambiguity, and keep going.
Input
Output
OpenAI GPT-5.5 Documentation
GPT-5.5 is OpenAI’s premiere frontier-tier model family, bringing unprecedented intelligence and enhanced adaptive reasoning to scale. Built to power ultra-long-horizon agentic automation and cross-disciplinary technical orchestrations, it sets a new baseline for structural execution fidelity, contextual cross-referencing, and complex software systems architecture without identity or performance degradation.
Key Capabilities
- Advanced State Refinement & Tracking: Natively maps out thousands of interdependent steps across multiple external tools and local system configurations, gracefully executing self-correction sequences when encountering runtime failures.
- Unified Architectural Alignment: Demonstrates profound comprehension across massive, multi-directory system frameworks, making it the definitive engine for cross-repository code base modernization, log aggregation, and system-wide security auditing.
- Extensive 400,000 Token Context Window: Natively houses massive application traces, historical transaction databases, or full configuration frameworks in a single interaction.
- Dynamic Reasoning Effort Scalability: Features a granular
reasoning_effortparameter (low,medium,high) allowing infrastructure developers to precisely match compute allocation with logical complexity. - Ultra-Expanded Output Target: Pushes maximum completion bounds to an expanded 128,000 token limit, ensuring continuous code blocks, architectural schematics, and structural documents compile without truncation.
Request Parameters
To interact with this model via the OpenAI Responses API, an OpenAI-compatible gateway, or local deployment routing:
| Parameter | Type | Required | Description |
|---|---|---|---|
model |
string |
Yes | Use "gpt-5.5". |
messages |
array |
Yes | Array of standard conversation objects supporting text, high-resolution visuals, and engineering payloads. |
reasoning_effort |
string |
No | Configures thinking depth budget. Supports "low", "medium", or "high". |
max_completion_tokens |
integer |
No | Limits maximum generation tokens. Supports up to 128,000 max output tokens. |
Reasoning Mode Configuration
gpt-5.5 supports internal reasoning via the reasoning_effort parameter. This controls how much compute the model dedicates to its hidden chain-of-thought before producing the final response.
Effort Levels
| Value | Description |
|---|---|
"low" |
Minimal reasoning; fastest and cheapest |
"medium" |
Balanced reasoning depth and speed |
"high" |
Maximum reasoning; best for hard multi-step problems |
Request Example
json
{
"model": "gpt-5.5",
"reasoning_effort": "high",
"messages": [
{
"role": "user",
"content": "Your question here"
}
]
}Note:
reasoning_effortis a top-level parameter in the request body. Not all GPT models support it — it applies to reasoning-capable variants only.
Unlock the most affordable AI hosting
Run models at scale with our fully managed GPU infrastructure, delivering enterprise-grade uptime at the industry's best rates.