# GPT 5.4

> GPT 5.4 is available on Artifio.ai — a unified AI workspace with a single pay-as-you-go wallet.

- Canonical page: https://artifio.ai/models/gpt-5-4
- Run it: https://artifio.ai/create/gpt-5-4
- Modality: prompt to text
- Model brand: OpenAI
- Pricing: from about $0.050 per generation (exact cost shown in the workspace before you confirm; varies with selected options)
- Typical generation time: ~25s (catalogue estimate)
- Full pricing table: https://artifio.ai/pricing/models

## About GPT 5.4

GPT-5.4 is a large language model OpenAI released on March 5, 2026, initially as GPT-5.4 Thinking and GPT-5.4 Pro for paid plans, with mini and nano variants following on March 17. Its defining addition is native computer use: the model can operate desktop software directly, and on the OSWorld-Verified benchmark it scored 75%, above both GPT-5.2's 47.3% and the 72.4% average recorded for humans. OpenAI also reported a 33% reduction in factual errors versus GPT-5.2.

Artifio offers GPT-5.4 as a text model in the same workspace as 100+ others, paid per generation from one wallet.

## Features

- 1,050,000-token context window with up to 128,000 output tokens.
- Native computer-use tooling alongside web search, file search, code interpreter, hosted shell, and MCP support in the API.
- Reasoning effort selectable from none through xhigh.
- Accepts text and image input; outputs text.
- Four variants: Thinking, Pro, mini (free-tier), and nano (API-only).
- Knowledge cutoff of August 31, 2025.

## What people use it for

- Agent workflows that drive real applications through screenshots, mouse, and keyboard.
- Deep research tasks, an area OpenAI specifically upgraded in this release.
- Coding and professional document work over very long context.
- Tool-heavy pipelines that benefit from the built-in tool search mechanism.

## Reported strengths

- State-of-the-art computer use at release: 75% on OSWorld-Verified, past the measured human baseline.
- 33% fewer factual errors than GPT-5.2, continuing the line's hallucination reductions.
- A context window over 1 million tokens, roughly 2.5x GPT-5.2's.
- Reception praised its token-efficient responses and low hallucination rate.

## Reported limitations

- Users reported that it sometimes does not follow prompts accurately, a criticism recorded in Wikipedia's reception summary.
- The mini and nano variants cost about four times as much through the API as their GPT-5 counterparts did, drawing pricing complaints.
- Prompts over 272,000 input tokens are billed at elevated rates for the whole session, so the million-token window carries a surcharge in practice.
- Its knowledge cutoff of August 31, 2025 sits half a year behind the release date.

## Sources

1. GPT-5.4 (Wikipedia): https://en.wikipedia.org/wiki/GPT-5.4
2. GPT-5.4 Model (OpenAI): https://developers.openai.com/api/docs/models/gpt-5.4

---

About Artifio.ai: 100+ AI image, video, and audio models in one workspace with a single shared wallet — no per-model subscriptions. This is the markdown rendition of https://artifio.ai/models/gpt-5-4 for AI assistants and crawlers; the HTML page is canonical.