OpenAI
OpenAI's GPT-6.2 Astra costs 20% less and Sol holds its price
OpenAI's GPT-6.2 Astra is listed at $8 per 1M input tokens, down from $10, and the API now drops reasoning items when a conversation prefix changes.

OpenAI added GPT-6.2 Astra and GPT-6.2 Sol to its API, with Astra priced 20% below GPT-6 Astra and Sol priced the same as GPT-6.1 Sol. The API also changes how it treats a model’s earlier reasoning, which can raise token use for apps that edit a conversation midway.
Both Models Are Available in the API
OpenAI listed gpt-6.2-astra and gpt-6.2-sol in its API changelog on Nov. 17, 2026. The entry reads: “Use GPT-6.2 Astra for demanding reasoning and coding tasks, or GPT-6.2 Sol for everyday coding and agent workflows.”
The models support the Responses API and Chat Completions. Tool calling requires the Responses API. They reached ChatGPT Work and Codex the same day, according to OpenAI’s ChatGPT and Codex changelog.
GPT-6.2 Astra is the first Astra model OpenAI has released since it decided not to release GPT-6.1 Astra. The BBC reported that decision Sept. 29, 2026.
Astra Costs 20% Less Than GPT-6 Astra
GPT-6.2 Astra is priced at $8 per 1M input tokens and $40 per 1M output tokens, according to OpenAI’s model page, read Nov. 17. GPT-6 Astra is listed at $10 and $50. Cached input drops 60%, from $1 to $0.40.
GPT-6.2 Sol costs $2 for input and $10 for output, with cached input at $0.10. Those match GPT-6.1 Sol’s listed prices.
A request over 272,000 input tokens is billed at double the input and cache rates and 1.5 times the output rate for the full request.
The API Now Drops Reasoning Items
OpenAI’s “Using GPT-6.2” guide, read Nov. 17, says the API will drop reasoning items when it detects a mid-conversation edit to the history (a prefix edit), a model downgrade or a change of end user or organization. Reasoning items are the record of the model’s earlier thinking, carried from one response to the next.
“When the model loses access to reasoning items, answer quality can decline and the model may use more tokens to repeat earlier work,” the guide says. It tells developers to keep the prefix unchanged when they change settings between responses, and to review integrations for edits “such as changes to instructions or tool calls.” An unchanged prefix also keeps earlier input eligible for prompt caching.
Sol Gets More Work Per Plan in Codex
OpenAI’s Codex changelog says that with GPT-6.2 Sol, “you can complete more tasks on the same paid plan than with GPT-6.1 Sol.” OpenAI’s Codex pricing page estimates 17 to 180 local messages per five-hour period on Plus and Standard Business for Sol, up from 15 to 160 on GPT-6.1 Sol. OpenAI describes the figures as estimates.
Our View
We think the reasoning-items change will decide Astra’s bill more than the price cut does. A developer whose application changes instructions or tool calls partway through a conversation now makes the API discard the model’s reasoning items.
OpenAI says the model may then spend more tokens redoing earlier work. Extra tokens are billed. A price per token only lowers a bill when the number of tokens stays the same, so a 20% cut on Astra can be spent on repeated work.
The same edit to the prefix also ends eligibility for prompt caching on earlier input, so the cached-input discount goes with it. Sol’s price did not move, and its answers can still decline when the reasoning items go missing. The Astra saving holds only for applications that keep the conversation prefix unchanged between responses.
