🚨 OpenAI just launched GPT-5.5. The OpenAI team was nice enough to give me early
By Allie K. Miller · April 24, 2026 · Curated by George's Blog
🚨 OpenAI just launched GPT-5.5.
The OpenAI team was nice enough to give me early access over the last several weeks, and I just want to flag: there is a certain class of models (one that we’re hitting now) where unless you are deep in code or scientific research, you might not even notice a difference.
The vast majority of business professionals may only notice that these models are concise as hell.
They’re direct in a way that may actually make you sit back.
They’re token efficient.
And while they might not be the greatest copywriters of all time, boy can they impact system connections.
OpenAI is calling the new model "a new class of intelligence for real work," with the biggest gains in coding, computer use, knowledge work, and research.
Agentic coding → 82.7% on Terminal-Bench 2.0 (up from 75.1% on 5.4, and ahead of Claude Opus 4.7 at 69.4%). Translation: it can plan, use tools, hold context, and finish multi-step technical work without hand-holding.
Knowledge work → 84.9% on GDPval, ahead of Claude Opus 4.7 and Gemini 3.1 Pro. 78.7% on OSWorld-Verified for operating real computer environments. OpenAI says 85% of their own company now uses Codex every week across finance, comms, marketing, and product. Their finance team reviewed 24,771 K-1 tax forms (71,637 pages) two weeks faster than last year - sheesh. A GTM employee automated weekly reporting and immediately got 5-10 hours a week back.
Scientific research → 85% on ARC-AGI-2 (up from 73.3%). 35.4% on FrontierMath Tier 4. A Jackson Lab researcher used it to analyze 62 samples and nearly 28,000 genes in work that would have taken his team months.
GPT-5.5 also matches GPT-5.4 per-token latency while performing at a much higher level of intelligence, and uses significantly fewer tokens to finish the same Codex tasks.
AVAILABILITY in ChatGPT: Plus, Pro, Business, Enterprise as of today. GPT-5.5 Pro for Pro, Business, Enterprise.
Docs and rollout info: https://lnkd.in/e4ySwSfc
I am expecting more data silos to break in the enterprise and more scientific breakthroughs in the next 2-3 years.