GPT-5.5-Pro

GPT-5.5 Pro belongs to the GPT-5.5 family, which OpenAI describes as the first fully retrained base model since GPT-4.5, rebuilt from the ground up rather than incrementally extended. The Pro tier sits on top of that lineage as the same underlying model operating with a parallel test-time compute setting, and OpenAI treats its offline safety results as strong proxies for the base while evaluating the Pro configuration separately where the setting could materially change the risk profile.

For practical use, GPT-5.5 Pro is positioned for complex real-world work such as writing code, researching online, analyzing information, producing documents and spreadsheets, and chaining tools to finish multi-step tasks with minimal prompting. OpenAI's GeneBench paper highlights the intended workload by reporting a 33.2% eval-level pass rate on multi-stage inference problems in genomics and quantitative biology, while also showing persistent gaps in which many problems remain under 20% pass rate and models often notice a diagnostic signal without acting on it. That combination of omnimodal input handling, agentic tool use, and test-time scaling makes GPT-5.5 Pro a fit for research and coding pipelines where the model can iterate, call tools, and keep working until a task is done, rather than for single-turn chat.

Poeopenai/gpt-5.5-progpt-pro

Open provider page Provider docs

Quick Info

Provider: Poe
Model key: openai/gpt-5.5-pro
Release date: Apr 8, 2026
Last updated: Apr 8, 2026
Knowledge cutoff: 2025-12-01
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $27.2727
Output token cost: $163.6364

Limits

Output tokens: 128,000 tokens
Context window: 400,000 tokens

Transparent token rates

Compare GPT-5.5-Pro pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about GPT-5.5-Pro

Poe

Official sourceAnnouncement

OpenAI announced GPT-5.5 on April 23, 2026, with an update on April 24 confirming that GPT-5.5 and GPT-5.5 Pro are now available in the API. The company describes GPT-5.5 as its "smartest and most intuitive to use model yet," designed to handle multi-part tasks by planning, using tools, checking its work, and navigatin OpenAI released GPT-5.5 Pro with its strongest safeguards to date, evaluated across its full safety and preparedness frameworks with internal and external red-teamers, plus targeted testing for advanced cybersecurity and biology capabilities, and feedback from nearly 200 early-access partners. The API deployment requir

Poe

Coverage

Techmeme's April 24, 2026 snapshot cites Bloomberg reporting that OpenAI launched GPT-5.5 on April 23, 2026, positioning it as a model designed to handle more complex tasks with minimal guidance, and stating that the model will power OpenAI's upcoming "super app." This confirms the official launch window for the GPT-5. For developers, the reported launch timing and minimal-guidance design intent suggest GPT-5.5 Pro is targeted at complex, agentic workloads rather than simple chat, aligning with the GeneBench evidence of improved multi-step inference capability. However, the Techmeme excerpt itself does not provide API-specific detail

Poe

CoverageBenchmark

OpenAI's GeneBench paper (dated April 23, 2026) introduces a benchmark of 103 multi-stage inference problems in genomics and quantitative biology, and explicitly reports results for a distinct GPT-5.5 Pro tier alongside the mainline GPT-5.5 family. The paper states that GPT-5.5 Pro reaches a 33.2% eval-level pass rate The GeneBench evaluation highlights a persistent capability gap even at the Pro tier: 60.2% and 62.1% of problems remain below 20% pass rate under repeated runs for the two strongest Pro-harness settings, and models often exhibit a "noticing vs. acting" failure mode, identifying diagnostic signals but failing to propag