OpenAI has released GPT-5.5, a fully retrained agentic model, just six weeks after GPT-5.4. The release is being pitched as the backbone of an AI “super app” — and benchmarks suggest the claim is more than marketing.
Benchmarks
- Terminal-Bench 2.0: 82.7% — state-of-the-art for complex command-line usage involving planning, iteration and tool coordination
- GDPval: 84.9%
- SWE-Bench Pro: 58.6% — real-world GitHub issue resolution
The Agentic Pitch
Instead of needing a human to babysit every step, GPT-5.5 is positioned to take on “messy, multi-part tasks” autonomously: writing and debugging code, researching online, analysing data, creating documents and spreadsheets, operating software, and moving across tools until a task is finished. It combines planning, tool use, self-checking, and ambiguity navigation into one model.
Availability
GPT-5.5 is rolling out to ChatGPT Plus, Pro, Business and Enterprise users, and is available in Codex. A stronger variant, GPT-5.5 Pro, is being released to Pro, Business and Enterprise tiers.
Why It Matters
As Fortune noted today, “AI model launches are starting to look like software updates.” The six-week cadence between GPT-5.4 and 5.5 — and the explicit agentic framing — suggest OpenAI is now shipping models in direct response to competitive pressure from Anthropic, Google and xAI, rather than on a legacy annual cycle. The Terminal-Bench 2.0 score in particular shows meaningful improvement over GPT-5.4.
Follow Vibes Uncut Media for AI model coverage.
Leave a Reply