# OpenAI's GPT-5.5 makes computer-use agents the default frontier
> Native state-of-the-art computer use, 1M-token context and Codex on Nvidia infra push agentic workflows mainstream

**Meta:** type: event · date: 2026-04-24 · heads: Comment la vie change, Le jeu long · 5 takes · 2 lenses · 1 regions

## Summary

[Openai](/fr/entity/openai) made GPT-5.5 (and GPT-5.5 Pro) its frontier model on 24 April 2026, building on March's GPT-5.4, the first general-purpose model with native, state-of-the-art computer-use, up to 1M-token context and long-horizon plan/execute/verify. GPT-5.5 powers Codex on [Nvidia](/fr/entity/nvidia) infrastructure; Codex now serves over 2m weekly users, up 5x in three months, and OpenAI APIs clear 15bn+ tokens/minute. The line collapses dedicated coding models into one agentic model that operates computers and works across spreadsheets, slides and documents. The capability arc tracks rivals: [Google's Antigravity agents](/fr/n/gemini-35-flash-agents-2026) and [Anthropic](/fr/entity/anthropic)'s Claude Code on Opus 4.8.

## By the numbers

- 24 Apr 2026, GPT-5.5 / 5.5 Pro available in the API.
- 1M tokens, context window.
- 2m+, weekly Codex users, up 5x in three months.
- 15bn+, tokens/minute across OpenAI APIs.

## Why it matters

Native computer-use moves the frontier from chat to action: an agent that can drive arbitrary software shifts AI from assistant to operator, with direct enterprise-workflow stakes. It also resets evaluation, capability now means task completion across tools, where [static coding benchmarks are degrading](/fr/n/swe-bench-audit-benchmark-rot-2026).

## What to watch

- Enterprise deployments of computer-use agents and the error/oversight regime.
- Safety incidents from agents acting across live systems.
- Whether GPT-5.6 ships before the IPO window.

## Regional takes (batched by bias / lens)

### unlabelled
- **OpenAI (Introducing GPT-5.5)** (United States, en) — OpenAI's own launch post for GPT-5.5, its frontier model as of June 2026, folding GPT-5.4's coding and computer-use advances into a single general-purpose model with long-horizon agentic capability.
  Source: https://openai.com/index/introducing-gpt-5-5/
- **OpenAI (Introducing GPT-5.4)** (United States, en) — The March GPT-5.4 post: first general-purpose model with native, state-of-the-art computer-use, up to 1M-token context, designed to plan/execute/verify across long horizons, the foundation GPT-5.5 built on.
  Source: https://openai.com/index/introducing-gpt-5-4/
- **OpenAI Developers (Codex changelog)** (United States, en) — OpenAI's Codex changelog tracking the model and agent updates behind the 2m+ weekly-user growth.
  Source: https://developers.openai.com/codex/changelog
- **OpenAI Help Center (model release notes)** (United States, en) — 
  Source: https://help.openai.com/en/articles/9624314-model-release-notes

### infrastructure / partner
- **NVIDIA Blog** (United States, en) — Nvidia's account of GPT-5.5 powering Codex on Nvidia infrastructure, framing the release as the agentic-coding workload that justifies large GPU clusters.
  > "OpenAI's new GPT-5.5 powers Codex on NVIDIA infrastructure for AI agents."
  Source: https://blogs.nvidia.com/blog/openai-codex-gpt-5-5-ai-agents/

## Across the graph
- Related: [[gemini-35-flash-agents-2026]], [[swe-bench-audit-benchmark-rot-2026]], [[openai-122bn-softbank-round-2026]]
- Entities: Openai, Nvidia, Microsoft

---
Canonical: https://rbtfl.xyz/fr/n/gpt-55-codex-computer-use-2026