OpenAI FRONTIER API · Context: 1.05M tokens (922K max input)

OpenAI GPT-6.2 Astra

OpenAI's Nov 2026 frontier model for long professional and agentic work. 20% cheaper per token than GPT-6 Astra with a 60% deeper cached-input discount, and fewer output tokens per task (vendor-reported). Drops reasoning items on mid-conversation prefix edits.

Provider
OpenAI (Proprietary API)
Architecture
Proprietary reasoning model (text + image in, text out); architecture undisclosed
Parameters
Proprietary Frontier
Context Window
1.05M tokens (922K max input) (Max Output: 128,000 tokens)
API Pricing
Input: $8.00 / 1M · Cached: $0.40 / 1M · Output: $40.00 / 1M
Hardware Target
Cloud API Only

Interactive Token Cost Estimator (GPT-6.2 Astra)

Est. API Cost (0% Cache): --/mo | With 70% Prompt Cache: --/mo
Compare all 12 models →

Architectural Strengths & Target Workloads

  • Agents' Last Exam: +5.2 pts over GPT-6 Astra at <1/3 the cost; +5.5 over Claude Fable 5.2 at ~5% cost (vendor-reported)
  • GDPval v2 +6.2 pts with 22% fewer output tokens vs GPT-6 Astra (vendor-reported)
  • $0.40 / 1M cached input for long agent sessions

Primary Engineering Workloads: Long-Horizon Agents · Professional Documents, Slides & Spreadsheets · Scientific / Terminal Workflows.

Frequently Asked Questions (GPT-6.2 Astra)

How much does OpenAI GPT-6.2 Astra cost per 1M tokens?

OpenAI GPT-6.2 Astra is priced at $8.00 per 1M uncached input tokens, $0.40 per 1M cached input tokens, and $40.00 per 1M output tokens (OpenAI API (gpt-6.2-astra), Responses/Chat Completions/Batch. Cache writes $10/1M (30 min) or $20/1M (12 h). >272K input: input/cache 2x, output 1.5x. Batch 50%, Fast 2x, Ultrafast 3x (6x from Mar 1, 2027). Also in ChatGPT Work, Codex and GitHub Copilot.).

What is the context window of OpenAI GPT-6.2 Astra?

It supports up to 1.05M tokens (922K max input) (1,050,000 tokens) and a maximum generation output of 128,000 tokens.

Full TechPulse Engineering Analysis

GPT-6.2 Astra and Sol: OpenAI Cuts Astra to $8 / $40, Keeps Sol at $2 / $10, and Changes How Reasoning Survives Prompt Edits

OpenAI's new pair targets long professional and agentic work with fewer output tokens per task. Both have a 1.05M-token window. For developers, the bigger news is in the migration guide: edit the conversation prefix mid-session and the API now drops the model's reasoning.

Read Full Technical Dispatch (5 min read) →
Risograph-style illustration of a tall pale tower with a slate-blue roof and a low terracotta-roofed pavilion standing on the same green ground under one ochre sun, joined by a thin ink line