OpenAIOpenAI·💬 Text Generation·New

GPT-6.2 Astra

ReasoningVisionFunction CallingWeb Searchanonymized
🧠 Try in Intelligence →Try on Venice.ai ↗
Quick reference
GPT-6.2 Astra — TLDR
  • 🏢 OpenAI's flagship GPT-6.2 model, released November 17, 2026.
  • 📏 1.05M-token context: 922K input, 128K output tokens.
  • 👁️ Text and image inputs; stronger visual reasoning for sites, slides, 3D.
  • 🔧 Function calling, web search, computer use, document work.
  • 🧠 Built for engineering, professional work, and scientific research.
  • ⚡ OpenAI reports better results using fewer output tokens than before.
  • 🎯 OpenAI reports 62.5% on Artificial Analysis's AA-AnalystAgent task set.
  • 💬 OpenAI says writing is more professional and readable than GPT-6 Astra.
💰 Pricing
$8.00 / $40.00
per 1M · input / output
📏 Context
1.1M tokens
📅 On Venice since
Nov 17, 2026
2 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 →
49 models on Venice
30 text · 6 image · 6 inpaint · 4 video · 2 embedding · 1 asr
Since Jan 15, 2025

About this model

GPT-6.2 Astra is OpenAI's frontier model for the hardest end-to-end projects, launched alongside the efficiency-oriented GPT-6.2 Sol and the higher-compute GPT-6.2 Astra Pro. Per OpenAI's API documentation, it can build and revise software, use a computer, and work with documents, accepting text and images and responding with text, with a 1.05M-token context window (922K input, 128K output) and support for function calling and web search.

Relative to its direct predecessor GPT-6 Astra, OpenAI positions 6.2 as delivering better results on demanding tasks while emitting fewer output tokens, so more work fits inside the same usage limits. OpenAI also says GPT-6.2 Astra's writing is more professional and readable than GPT-6 Astra's, and highlights stronger visual reasoning applied to websites, slideshows, and 3D designs.

On evaluations cited in OpenAI's own launch post, GPT-6.2 Astra answers 62.5% of questions correctly on all five independent attempts in Artificial Analysis's AA-AnalystAgent suite, which works on real spreadsheets and reports, and OpenAI reports a new high for the model on Terminal-Bench-Science 0.1, a benchmark for scientific workflows carried out in a terminal. These figures are self-reported by the provider.

For lineage context: GPT-6 Astra, released in September 2026, was already OpenAI's flagship for complex reasoning, coding, computer use, research, and document creation, with configurable reasoning effort levels. GPT-6.2 Astra continues that line, with long-context requests above 272K input tokens billed at higher rates.

This About section is AI-generated from public sources (Claude Opus 5), 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