How long does the GPT-5.6 Sol price cut last?

OpenAI set the lower GPT-5.6 Sol rates to run for three months, from August 21, 2026 through November 21, 2026. OpenAI has not said what the price will be after that window closes. Teams budgeting past November should treat these rates as promotional rather than as the new list price.

Does the GPT-5.6 Sol price cut change my ChatGPT subscription?

No. The GPT-5.6 Sol reduction covers API usage plus credits on ChatGPT Work and Codex for eligible plans. OpenAI confirmed that Pro, Plus and Business subscription prices stay the same. If you reach GPT-5.6 Sol through one of those plans rather than the API, your monthly bill does not change.

Which OpenAI models were discounted before GPT-5.6 Sol?

OpenAI cut prices on two smaller GPT-5.6 models shortly before the Sol reduction: GPT-5.6 Terra came down 20% and GPT-5.6 Luna came down 80%, according to Reuters. The GPT-5.6 Sol cut extends the same pattern up to OpenAI's frontier tier, which had stayed at full list price until now.

Does the discount apply to cached input tokens?

Yes. GPT-5.6 Sol cached input falls alongside the standard input rate, so prompt caching still costs a tenth of fresh input. Workloads that reuse a long system prompt or a large document across many calls compound both savings — the caching discount and the new lower base rate.

OpenAI · 2026-08-21 · major

GPT-5.6 Sol price cut — OpenAI drops API rates over 20% for three months

OpenAI cut GPT-5.6 Sol API pricing by over 20% for three months. Input falls from $5 to $4 per million tokens, cached input from $0.50 to $0.40, and output from $30 to $20. Codex credits and ChatGPT Work get the same rates.

OpenAI Developers API Docs pricing banner on a pink and orange gradient

OpenAI's frontier GPT-5.6 Sol now costs over 20% less to call, for the next three months.

Quick facts

Maker	OpenAI
Model	GPT-5.6 Sol
Size of cut	Over 20%
Announced	August 21, 2026
Runs through	November 21, 2026
Where it applies	API, Codex credits, ChatGPT Work
Not affected	Pro, Plus and Business plans

Pricing

Input · Was $5.00	$4.00 / 1M tokens
Cached input · Was $0.50	$0.40 / 1M tokens
Output · Was $30.00	$20.00 / 1M tokens

source ↗

What is it?

GPT-5.6 Sol just got cheaper to run. OpenAI cut its API and credit pricing by over 20% on August 21, 2026. The lower rates are promotional, not permanent: they run for three months and end on November 21, 2026. Sol is OpenAI's frontier model for long-running agent and coding work.

How does it work?

The new rates cover standard short-context API calls, and the same discount flows into Codex credits and ChatGPT Work for eligible plans. Cached input drops by the same share as fresh input, so GPT-5.6 Sol keeps its usual 10x prompt-caching advantage on top of the lower base rate. Subscription pricing is untouched — Pro, Plus and Business members pay what they paid before.

Why does it matter?

Output tokens are where agent workloads spend most of their budget, and Sol's output rate takes the deepest cut in the table — a third off. Teams running long agent loops, large code reviews or high-volume batch jobs see the biggest saving. Reuters reports the cut follows earlier reductions on GPT-5.6 Terra and GPT-5.6 Luna, and ties both to competition from Anthropic and from Chinese AI providers.

Who is it for?

teams running high-volume agent and coding workloads

Frequently asked questions

How long does the GPT-5.6 Sol price cut last?: OpenAI set the lower GPT-5.6 Sol rates to run for three months, from August 21, 2026 through November 21, 2026. OpenAI has not said what the price will be after that window closes. Teams budgeting past November should treat these rates as promotional rather than as the new list price.
Does the GPT-5.6 Sol price cut change my ChatGPT subscription?: No. The GPT-5.6 Sol reduction covers API usage plus credits on ChatGPT Work and Codex for eligible plans. OpenAI confirmed that Pro, Plus and Business subscription prices stay the same. If you reach GPT-5.6 Sol through one of those plans rather than the API, your monthly bill does not change.
Which OpenAI models were discounted before GPT-5.6 Sol?: OpenAI cut prices on two smaller GPT-5.6 models shortly before the Sol reduction: GPT-5.6 Terra came down 20% and GPT-5.6 Luna came down 80%, according to Reuters. The GPT-5.6 Sol cut extends the same pattern up to OpenAI's frontier tier, which had stayed at full list price until now.
Does the discount apply to cached input tokens?: Yes. GPT-5.6 Sol cached input falls alongside the standard input rate, so prompt caching still costs a tenth of fresh input. Workloads that reuse a long system prompt or a large document across many calls compound both savings — the caching discount and the new lower base rate.

Try it

Model id gpt-5.6-sol on the OpenAI API