GPT-6 Astra pricing: the API costs twice as much, and your ChatGPT bill did not change
GPT-6 Astra is $10 per million input tokens and $50 output, against $5 and $30 for GPT-5.6 Sol. Twice the price on input, two thirds more on output, and nothing changed for ChatGPT subscribers.
Verdict
GPT-6 Astra costs $10 per million input tokens and $50 output, against $5 and $30 for GPT-5.6 Sol. It is now in Plus, Pro, Business and Enterprise, and no plan price changed.
The number that matters#
2x input
GPT-6 Astra API cost against GPT-5.6 Sol
$10 and $50 per million, against $5 and $30 for GPT-5.6 Sol. Output rises by two thirds.
If you have anything running on the OpenAI API and you swap the model string without changing anything else, your bill roughly doubles. That is the whole story for most small businesses building on it.
| Per million tokens | GPT-5.6 Sol | GPT-6 Astra | Change |
|---|---|---|---|
| Input | $5 | $10 | 2x |
| Output | $30 | $50 | 1.67x |
Where you land between those two depends on your input-to-output ratio. Prompt-heavy work with short answers sits nearer 2x; work that generates long output sits nearer 1.67x.
Full API pricing#
| Per million tokens | |
|---|---|
| Input | $10.00 |
| Output | $50.00 |
| Cached input | $1.00 |
| Cache write | $12.50 |
Above 272,000 input tokens, a higher rate applies. So the million-token context window is available, but it is not priced flat across its whole range.
What you get for the money#
| GPT-6 Astra | |
|---|---|
| Context window | 1,050,000 tokens |
| Max output | 128,000 tokens |
| Knowledge cutoff | 30 April 2026 |
| Model ID | gpt-6-astra |
The million-token context is the headline capability. Work that previously needed splitting a long document into chunks, summarising each, and stitching the summaries back together can now go in whole.
Whether that is worth double depends entirely on whether you were doing that chunking. If you were, the engineering you delete may cost more than the tokens you add. If your prompts are a few thousand tokens of ordinary text, you are paying double for a window you will never fill.
Nothing changed for ChatGPT subscribers#
This is the part most coverage buried.
| Plan | Price before | Price now |
|---|---|---|
| Free | $0 | $0 |
| Go | $8/mo | $8/mo |
| Plus | $20/mo | $20/mo |
| Pro Codex | $100/mo | $100/mo |
| Pro Max | $200/mo | $200/mo |
| Business | $20/user/mo annually | $20/user/mo annually |
There is no standalone GPT-6 subscription. There is no new tier. Nothing on your card changed.
The staged rollout OpenAI announced on 3 September has completed. As of 8 September:
| Plan | GPT-6 Astra |
|---|---|
| Free | No |
| Go | No |
| Plus | Yes |
| Pro | Yes, expanded access |
| Business, Enterprise | Yes |
So the cheapest subscription that reaches it is Plus at $20 a month, unchanged from before GPT-6 existed. Pro at $100 and $200 gets expanded access rather than a different model.
Full plan breakdown: ChatGPT pricing.
Should you switch?#
| If you | Do this |
|---|---|
| Have nothing on the API | Nothing. This launch does not touch you |
| Run a working app on GPT-5.6 | Stay. Test GPT-6 on a copy first, and only if quality is your bottleneck |
| Chunk long documents to fit a context window | Test it properly. Deleting that pipeline may be worth the price |
| Are cost-sensitive and shipping fine | Stay, and set up prompt caching regardless |
The reflex to move to the newest model on release day is expensive and rarely justified. A doubled input price needs a reason, and "it is newer" is not one.
The short version
What works
- A 1,050,000-token context window is a genuine step up, and long-document work that needed chunking may no longer need it
- Cached input at $1 per million is a tenth of the standard rate, so repeated system prompts stay cheap
- ChatGPT subscription prices did not move, so if you are not on the API this launch costs you nothing
What does not
- The API is twice the input price of the model most teams are already using, and two thirds more on output
- Prompts above 272,000 input tokens are billed at a higher rate, so the large context window is not uniformly priced
- Free and Go do not get it, so the cheapest way in is Plus at $20 a month
Frequently asked questions
- How much does GPT-6 Astra cost?
- Through the API, $10 per million input tokens and $50 per million output tokens. Cached input is $1 per million and cache writes are $12.50 per million. Prompts above 272,000 input tokens are billed at a higher rate.
- Is GPT-6 more expensive than GPT-5.6?
- Yes. GPT-5.6 Sol is $5 per million input and $30 output, against $10 and $50 for GPT-6 Astra. That is exactly double on input and about two thirds more on output, so a swap roughly doubles the bill depending on how your workload splits between the two.
- Did ChatGPT get more expensive because of GPT-6?
- No. Free, Go, Plus, Pro and Business all cost exactly what they cost before the announcement. The API price rise does not pass through to subscription plans.
- Can I use GPT-6 Astra in ChatGPT?
- Yes, if you are on Plus, Pro, Business or Enterprise. The staged rollout that began on 3 September 2026 has completed for those plans, and Pro carries expanded access rather than standard. Free and Go do not include it, so Plus at $20 a month is the cheapest way to reach it through a subscription.
- What is the GPT-6 Astra context window?
- 1,050,000 tokens of input with a maximum output of 128,000 tokens. The knowledge cutoff is 30 April 2026 and the model ID is gpt-6-astra.
Written by
Tashawar Awais
Researcher and editor
Tashawar handles verification and editing. Every figure in a review is checked a second time before it goes out, and anything that cannot be traced to a vendor page or a documented source is either qualified or cut. Where pricing is genuinely unclear, as it is with Canva team plans or Close CRM tiers, the article says so and tells the reader to confirm directly instead of quoting a number with false confidence.
- Second check on every published figure
- Removes claims the sources do not support
- Flags pricing that changes without notice