OpenAI GPT-6 Astra Is Now Available

OpenAI has made GPT-6 Astra available through its API, and it is now in the AI Data Stream model picker as the top-tier premium OpenAI option. Nothing has been retired: GPT-5.6 Sol, Terra and Luna all remain, so you can compare Astra on your own analytics before deciding whether it earns its price.

Pricing

Astra is priced above the rest of the OpenAI lineup:

• Input: $10 per million tokens
• Cached input: $1 per million tokens
• Cache writes: $12.50 per million tokens
• Output: $50 per million tokens

That is 2.5 times GPT-5.6 Sol ($4 / $20, OpenAI’s promotional price since August 21) and five times GPT-5.6 Terra’s input price ($2 / $12). Two details matter for the way this app works:

• Caching softens the cost of long conversations. Every follow-up question re-reads the conversation so far, and OpenAI bills that repeated prefix at the cached rate - a tenth of the normal input price. The first time a prefix is cached it is billed as a cache write at 1.25x the input rate ($12.50); every later read of it costs $1. Caching is automatic; there is nothing to switch on.
• Very large requests cost more. OpenAI bills requests with more than 272K input tokens at 2x the input rate and 1.5x the output rate. Normal analytics conversations sit far below that, but it is worth knowing before pasting in very large documents.

As always, AI Data Stream is bring-your-own-key: you are billed by OpenAI directly at their published rates.

What’s New

OpenAI describes Astra as its most capable model, built for demanding end-to-end professional work, with a context window of just over one million tokens. The claims most relevant to analytics come from the companies that tested it before launch: Databricks reports “a clear step up on data reasoning and document understanding” along with a better cost per task than GPT-5.6 Sol, and Box found it more than 10% less likely to make confidently incorrect assertions - it was better at declining to state conclusions the source material did not support. For a tool whose job is to tell you what your data does and does not show, that second point is the one to test. Both quotes are from OpenAI’s post on Astra for professional work; see OpenAI’s announcement for the full details.

GPT-5.6 Has Been Kept

Nothing has been retired and nothing has been migrated. GPT-5.6 Sol, Terra and Luna stay in the picker exactly as they were, and if one of them is your default it stays your default - no action needed. Astra is an opt-in choice for the questions where you want the strongest model OpenAI offers.

Reasoning Effort Control

Astra always reasons before it answers, and you control how much:

• Low - Faster responses, lower cost
• Medium - Balanced speed and depth
• High - Thorough reasoning for complex analysis
• Extra high - Deeper reasoning for difficult, multi-step questions
• Max - Maximum reasoning for the hardest questions

Alongside the Astra launch we have improved effort control across all OpenAI models:

• Extra high and Max are now available on GPT-5.6 Sol, Terra and Luna as well as Astra. GPT-5 Mini continues to offer low, medium and high.
• Your team default is now respected in chat. The effort level you save for a provider in Settings > AI Providers is the level the chat starts on when you pick that provider. You can still override it per message from the effort dropdown.
• Effort now applies fully when the model is using your data tools. We have moved our OpenAI integration to OpenAI’s newer Responses API, which Astra requires for tool use. On the previous API, GPT-5.6 requests that used data tools could fail at most effort levels; they now work at every level, and the model’s reasoning carries through from one tool call to the next within an answer.

One thing that has not changed: we send every request with storage disabled, so OpenAI does not keep a retrievable copy of your conversations on its side. Your conversation history lives in your AI Data Stream account.

The Full OpenAI Lineup

• GPT-6 Astra - Most powerful, premium; for the hardest analysis
• GPT-5.6 Sol - Most capable of the GPT-5.6 family
• GPT-5.6 Terra - Balanced performance and price
• GPT-5.6 Luna - Fast and cost-efficient
• GPT-5 Mini - Fast, budget-friendly
• GPT-4o - Reliable previous generation
• GPT-4o Mini - Budget-friendly option

Switching Models

1. Go to Settings > AI Providers
2. Click your OpenAI provider
3. Select your preferred model from the dropdown
4. Save

Or simply select it from the model picker when starting a new conversation.

Questions?

If you have any questions, contact us or check our AI providers documentation.