OpenAI dropped GPT-5.5 on April 23, 2026 — and the pitch is simple: a smarter model that doesn't ask you to hold its hand. OpenAI describes it as their "smartest and most intuitive to use model yet," one that understands what you're trying to do faster and can carry more of the work itself. That framing isn't just marketing — it reflects a genuine architectural shift toward autonomy.
The launch comes less than two months after OpenAI released GPT-5.4, the latest sign of the breakneck pace of development driving the AI sector. Seven weeks between major model releases is now the tempo, and OpenAI shows no signs of slowing down.
How GPT-5.5 Works
At its core, GPT-5.5 is built for agentic operation. 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. Relative to earlier models, GPT-5.5 understands the task earlier, asks for less guidance, uses tools more effectively, checks its work, and keeps going until it's done.
GPT-5.5 was co-designed for, trained with, and served on NVIDIA GB200 and GB300 NVL72 systems — a hardware partnership that directly enabled the efficiency gains OpenAI is touting. It supports a 1M token context window, image input, structured outputs, function calling, prompt caching, built-in computer use, hosted shell, MCP, and web search.
Key Performance Benchmarks
The numbers are hard to ignore. The model achieves 82.7% on Terminal-Bench 2.0 — a benchmark testing complex command-line workflows — beating Claude Opus 4.7 at 69.4% and Gemini 3.1 Pro at 68.5%. On GDPval, which tests agents' abilities to produce well-specified knowledge work across 44 occupations, GPT-5.5 scores 84.9%. On OSWorld-Verified, which measures whether a model can operate real computer environments on its own, it reaches 78.7%. And on Tau2-bench Telecom, which tests complex customer-service workflows, it reaches 98.0% without prompt tuning.
In scientific research, the model's scientific capabilities are now strong enough to meaningfully accelerate progress at the frontiers of biomedical research as a bona fide co-scientist. GPT-5.5 scores 80.5% on BixBench, a bioinformatics and data analysis evaluation, up from 74.0% for GPT-5.4, and 25.0% on GeneBench, a new evaluation focused on multi-stage scientific data analysis in genetics and quantitative biology.
When it comes to models the general public can access, GPT-5.5 has retaken the crown for OpenAI, achieving state-of-the-art across 14 benchmarks compared to 4 for Claude Opus 4.7 and 2 for Google Gemini 3.1 Pro.
Pricing
At $5 per million input tokens and $30 per million output tokens, GPT-5.5 is priced above GPT-5.4. GPT-5.5 Pro is considerably more expensive at $30 per million input tokens and $180 per million output tokens. OpenAI CEO Sam Altman argued on X that token efficiency gains offset the cost — GPT-5.5 completes the same Codex tasks with fewer tokens, which means cheaper runs even at a higher per-token rate.
