OpenAI released GPT-5.5 on April 23, codenamed “Spud,” positioning the model as its most capable system for autonomous, multi-step work.
The launch arrives one week after Anthropic shipped Claude Opus 4.7, setting up a direct comparison between the two frontier models.
GPT-5.5 Targets Agentic Work and Coding
GPT-5.5 is built to plan, execute, verify, and iterate across tools without constant human oversight. OpenAI describes it as “a new class of intelligence for real work and powering agents.”
Follow us on X to get the latest news as it happens
“We believe in iterative deployment; although GPT-5.5 is already a smart model, we expect rapid improvements. Iterative deployment is a big part of our safety strategy; we believe the world will be best equipped to win at the team sport of AI resilience this way,” wrote Sam Altman in a post.
The model rolls out now to ChatGPT Plus, Pro, Business, and Enterprise users. A more powerful Pro variant is also available. API pricing starts at $5 per million input tokens and $30 per million output tokens with a one-million-token context window.
OpenAI’s own benchmarks show GPT-5.5 ahead of Claude Opus 4.7 on several agentic tasks. It scored 82.7% on Terminal-Bench 2.0, compared to 69.4% for Opus 4.7.
On FrontierMath Tiers 1 through 3, it reached 51.7% versus 43.8%. Early independent tests reported similar trends in coding and knowledge-work evaluations.
Where Claude Opus 4.7 Still Leads
Anthropic’s model retains advantages in research writing, legal and financial reasoning, and instruction-following consistency, according to independent reviewers.
Opus 4.7 also supports higher-resolution vision at up to 3.75 megapixels, more than three times its predecessor.
On computer use, the gap narrows. GPT-5.5 scored 78.7% on OSWorld-Verified, while Opus 4.7 posted 78.0%.
The two models also trade leads on browsing benchmarks, with GPT-5.5 Pro pulling ahead at 90.1% versus 79.3%.
AI Race Accelerates in 2026
The back-to-back launches reflect a broader pattern. OpenAI has shipped multiple GPT-5.x variants this year, while Anthropic has steadily upgraded Claude through successive releases.
Google’s Gemini 3.1 Pro also competes for the same enterprise market.
For developers choosing between the two, the decision may come down to use case. GPT-5.5 appears stronger for agentic automation and long-horizon coding.
Claude Opus 4.7 may suit precision-heavy analytical workflows better. Whether independent benchmarks confirm OpenAI’s published numbers will become clearer in the days ahead.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights
