MODEL

GPT-5.5

OpenAI's current frontier flagship — released April 2026

Shipped by OpenAI · · class: frontier-llm

OpenAI released GPT-5.5 on April 23, 2026, with the GPT-5.5 Pro variant becoming available in the API on April 24. OpenAI framed the model as its “smartest and most intuitive” release to date, with stronger end-to-end task autonomy — writing and debugging code, online research, document and spreadsheet creation, software operation across tools without re-prompting. GPT-5.1 (Instant / Thinking / Pro) was retired from ChatGPT in March 2026 to make room.

What was claimed at release: faster intent inference; sustained multi-step task completion across heterogeneous tools; meaningful coding benchmark gains over GPT-5 and 5.1.

What’s been observed since: most reviewers characterize the deltas vs GPT-5 as iterative, with the autonomy gains real but bounded (long-horizon tasks still degrade past ~10–15 tool calls). The bigger story is the pace: GPT-5 → 5.1 → 5.5 inside ~6 months is the fastest cadence OpenAI has run on a flagship line, raising questions about whether the version numbers correspond to capability ticks or marketing rhythm.