OpenAIOpenAI·💬 Text Generation

GPT-5.4

ReasoningVisionFunction CallingWeb Searchanonymized
🧠 Try in Intelligence →Try on Venice.ai ↗
Quick reference
GPT-5.4 — TLDR
  • 🆕 OpenAI frontier model unifying the GPT and Codex lines.
  • 📏 Supports up to a 1M+ token context window.
  • 🧠 Adaptive reasoning effort: none, low, medium, high, xhigh.
  • 🔧 Native computer use plus deferred tool-loading tool search.
  • 👁️ Multimodal: accepts text and image inputs.
  • 📚 Inherits GPT-5.3-Codex coding into the mainline model.
  • ⚡ Native compaction extends longer agent trajectories.
  • 🌐 Improved deep web research and long-context handling.
💰 Pricing
$3.13 / $18.80
per 1M · input / output
📏 Context
1M tokens
📅 On Venice since
Mar 5, 2026
91 days ago
Provider

OpenAI is an American artificial intelligence research organization headquartered in San Francisco, structured as both a for-profit public benefit corporation and a nonprofit foundation. The lab developed the GPT family of large language models, the DALL-E…

Read full profile →
24 models on Venice
13 text · 4 video · 2 image · 2 embedding · 2 inpaint · 1 asr
Since Jan 15, 2025

About this model

GPT-5.4, released in March 2026, is OpenAI's frontier model aimed at complex professional work, folding together advances in reasoning, coding, and agentic workflows into a single system. A defining change is its very large context window of roughly 1.05M tokens, which OpenAI says makes it easier to analyze entire codebases, long document collections, or extended agent trajectories in one request. It accepts text and image inputs and exposes a reasoning-effort control ranging from none through xhigh.

Compared with the earlier GPT-5.2, OpenAI reports improvements in coding, document understanding, tool use, and instruction following. The model also incorporates the coding capabilities of GPT-5.3 Codex into the mainline line, and is described as OpenAI's first general-purpose model with native, built-in computer-use capabilities, enabling agents to operate software in a build-run-verify-fix loop. Two further additions—native compaction for longer agent runs and tool search via deferred tool loading—target larger tool ecosystems and reduced token usage.

GPT-5.4 ships alongside lighter and heavier variants, including GPT-5.4 Mini and GPT-5.4 Pro, the latter also offering the 1M+ context window. It was later succeeded within the same family by GPT-5.5. For teams building agents, automating document-heavy workflows, or handling multi-step tool use, GPT-5.4 represents OpenAI's consolidated frontier endpoint.

This About section is AI-generated from public sources (Claude Opus 4.8), with no human editing. It may contain inaccuracies — verify critical details against the sources listed above.

Data sources: Venice API · HuggingFace · Wikipedia — enrichment updated 1d ago