GPT-6 Astra API Pricing: $10 per Million Input Tokens, $50 Output
OpenAI released GPT-6 Astra in its API on September 3, 2026, describing it in the API changelog as its most capable model, built for reasoning, coding, computer use, research and document creation. It is available through the Responses and Chat Completions endpoints.
Prices per 1M tokens
From OpenAI’s pricing page, checked September 16, 2026:
| Tier | Prompt size | Input | Cached input | Cache writes | Output |
|---|---|---|---|---|---|
| Standard | up to 272K | $10.00 | $1.00 | $12.50 | $50.00 |
| Standard | over 272K | $20.00 | $2.00 | $25.00 | $75.00 |
| Batch and Flex | up to 272K | $5.00 | $0.50 | $6.25 | $25.00 |
| Batch and Flex | over 272K | $10.00 | $1.00 | $12.50 | $37.50 |
| Fast mode | up to 272K | $20.00 | $2.00 | $25.00 | $100.00 |
| Fast mode | over 272K | $40.00 | $4.00 | $50.00 | $150.00 |
A request with 1,000 input and 500 output tokens costs about $0.035 at standard rates. Fast mode is not available for GPT-6 Astra with EU data residency, and regional processing endpoints add 10% to the price.
How it compares
GPT-6 Astra costs 2.5 times as much as GPT-5.6 Sol on input ($4) and 2.5 times as much on output ($20), although Sol’s price is a promotion that runs at least through November 21, 2026. Among other vendors, Anthropic’s Claude Fable 5.1 lists the same $10 / $50 standard rates, and Claude Opus 5 costs half that at $5 / $25.
Output is where the bill grows. GPT-6 Astra is a reasoning model, and OpenAI bills thinking tokens as output, so long agent runs cost far more than the per-request figure above.
Changes to plan for when migrating
OpenAI lists several differences from earlier models:
- The
nonereasoning effort level is not supported. - Custom
temperature,top_pand log probabilities are not supported. - Tool calling requires the Responses API; Chat Completions users with tools need to migrate.
- Misalignment monitoring checks supported Responses API requests in the background and can raise safety alerts or stop a conversation for review.
On the same day OpenAI added async tool calling and mid-turn steering over WebSockets to the Responses API for long-running GPT-6 Astra work.
Compare GPT-6 Astra with every other OpenAI model on our OpenAI API pricing page, or across vendors in the LLM API pricing table.