## On April 23, 2026, OpenAI officially released GPT-5.5, a massive leap forward that marks the first fully retrained base model since GPT-4.5. Internally codenamed "Spud," the model isn't just an incremental update; it represents a fundamental shift toward truly agentic, "set-and-forget" artificial intelligence.
### The "Messy Prompt" Solution
The standout feature of GPT-5.5 is its ability to navigate ambiguity. OpenAI claims the model is designed for complex, real-world work—tasks where the user doesn't have a clear step-by-step plan.
*Autonomy:** Instead of requiring constant prompting, GPT-5.5 can take a "messy," multi-part request and independently plan, use tools, check its own work, and navigate through unclear instructions until the task is complete.
*Efficiency:** Despite the jump in intelligence, GPT-5.5 matches the latency of GPT-5.4. Remarkably, it uses significantly fewer tokens to complete the same tasks in Codex, making it both smarter and more cost-effective for developers.
### Agentic Coding and "Codex" Integration
The launch was held in close partnership with NVIDIA, utilizing their GB200 and GB300 NVL72 rack-scale systems.
*Codex Evolution:** GPT-5.5 now powers the new version of Codex, enabling "agentic coding" where the AI manages entire repositories rather than just snippets.
*Widespread Adoption:** Over 10,000 NVIDIA employees were granted early access, with engineers describing the model’s ability to solve complex information problems as "mind-blowing."
### Safety and the "Pro" Tier
OpenAI has launched two versions: GPT-5.5 and GPT-5.5 Pro.
*Pro Version:** Designed for extreme reasoning tasks, using parallel "test-time compute" to think deeper before responding.
*Safeguards:** This release includes the most robust safety suite to date, featuring specialized red-teaming for advanced cybersecurity and biology capabilities.
*Context Window:** The model features a 400K context window for Codex users, with plans to expand to 1M tokens for API deployments.
### Availability and Pricing
GPT-5.5 is currently rolling out to Plus, Team, and Enterprise subscribers.
*API Access:** Expected "very soon," with pricing set at $5 per 1M input tokens and $30 per 1M output tokens (roughly double the cost of GPT-5.4).
*Pro Pricing:** The high-tier Pro model will carry a premium price of $30/$180 per million tokens.
> "GPT-5.5 marks a genuine turning point. We've moved from a model that answers questions to a model that gets computer work done." —
#AaveAnnouncesDeFiUnitedReliefFund