GPT-6 Astra Context Window
Astra supports 1,050,000 tokens of context and up to 128,000 output tokens — with a pricing threshold you should know before sending huge prompts.
1,050,000 tokens total context · 128,000 tokens maximum output.
What 1.05 million tokens enables
The large context window is designed for repository-scale code, research collections, multi-document analysis and agent workflows that need substantial working context. In practice, usable capacity still depends on output reservation, tool traffic and how efficiently the prompt is structured.
Long context is not priced like ordinary context
OpenAI states that prompts above 272,000 input tokens use 2× input/cache pricing and 1.5× output pricing for the full request. That makes retrieval, chunking and prompt selection economically important even when the model can technically hold the entire corpus.
Official long-context results
| OpenAI MRCR v2 | GPT-6 Astra | GPT-5.6 Sol |
|---|---|---|
| 8-needle, 256K–512K | 100.0% | 91.5% |
| 8-needle, 512K–1M | 96.3% | 73.8% |
These are vendor-reported evaluation results. They are useful signals, not a guarantee for arbitrary production documents.
Frequently asked questions
What is the GPT-6 Astra context window?
OpenAI lists a 1,050,000-token context window.
What is GPT-6 Astra’s maximum output?
The listed maximum output is 128,000 tokens.
Does a 1M-token prompt cost more?
Yes. OpenAI applies long-context multipliers once input exceeds 272K tokens.
Primary sources
Specifications and prices can change. We prefer first-party documentation and record the date each page was reviewed.