Model launchAI modelsNews from
OpenAI releases GPT-6 Astra at $10 input and $50 output per 1M tokens
GPT-6 Astra reached OpenAI's API on September 3, 2026. It costs $10 input and $50 output per 1M tokens and has a 1,050,000-token context window.
OpenAI released GPT-6 Astra (gpt-6-astra) in its API on September 3, 2026. The changelog calls it
OpenAI’s most capable model, built for the hardest end-to-end work. It is also the most expensive
model in the GPT-6 family.
What it costs
Standard prices per 1M tokens, from OpenAI’s pricing page on October 18, 2026:
| Prompt size | Input | Cached input | Output |
|---|---|---|---|
| Up to 272K tokens | $10.00 | $1.00 | $50.00 |
| Over 272K tokens | $20.00 | $2.00 | $75.00 |
- Cache writes cost $12.50 per 1M tokens.
- Above 272K input tokens, the higher rates apply to the whole request, not just the extra tokens.
- Batch and Flex cost half. Fast mode (the new name for Priority processing since July 30, 2026) costs double.
- On September 29, OpenAI added an Ultrafast mode for Astra, with US data residency only, at $60 input and $300 output per 1M tokens.
Anthropic’s top general model, Claude Fable 5.1, lists the same $10 input and $50 output.
Specs
| GPT-6 Astra | |
|---|---|
| Context window | 1,050,000 tokens (max input 922,000) |
| Max output | 128,000 tokens |
| Knowledge cutoff | April 30, 2026 |
| Input / output | Text and images in, text out |
| Reasoning effort | low, medium, high, xhigh, max |
| Endpoints | Responses, Chat Completions, Batch |
Rate limits run from 5,000 requests and 1M tokens a minute on the Build tier to 15,000 requests and 40M tokens a minute on Grow.
Changes for developers
The changelog lists what Astra drops: there is no none reasoning effort, no custom temperature or
top_p, and no log probabilities. Tool calling needs the Responses API. OpenAI also runs
misalignment monitoring on supported Responses API requests, which can raise alerts or stop a
conversation for review.
In ChatGPT
OpenAI’s announcement, posted on its developer forum, says Astra went first to Pro, Enterprise and Business Premium users in ChatGPT Work and Codex, with Plus and Business users to follow within a few days.
What to check
Astra’s output costs five times as much as GPT-6.1 Sol’s. Before you switch a workload, run your own prompts on both and see whether Sol is good enough.
- GPT-6 Astra$50
- GPT-6.1 Sol$10
- GPT-6 Sol$10
- GPT-6 Luna$0.50
Sources
Read on . Prices are in US dollars. How we check facts