Four ways to pay for a model, and one Critical rating
GPT-6 Astra, and The Deploy Log for this week. Every price and claim read off the publisher's own page, with a call on each.
OpenAI
GPT-6 Astra
WAITif it touches customers
WHAT SHIPPEDGPT-6 Astra, rolling out to the OpenAI API and through Microsoft Azure and Amazon Bedrock at $10 per million input tokens and $50 per million output, the same headline price as Anthropic's Claude Fable 5.1, which shipped the same week. Cache reads are priced separately, and Fast mode costs 2x.
THE CATCHOpenAI rates Astra Critical on cyber capability under its own Preparedness Framework. Its announcement says Astra "is a significant jump in cyber capabilities and meets the Critical threshold in cybersecurity", and that its ability to identify and develop zero-day exploits "creates a need for stronger safeguards". It ships anyway.
IF YOU RUNa customer-facing surface, keep Astra off it until you can measure what it does. Run it internally, build the evaluation first, and read a vendor writing Critical about its own model as the warning it is.
SOURCEOpenAI, GPT-6 Astra: A new generation of intelligence
The Deploy Log
Models and APIs
| DEPLOYif you run long agentic contexts Claude Fable 5.1. Unchanged at $10 per million input and $50 per million output, with cache reads down to $0.25 per million, and Mythos 5.1 alongside it. Anthropic |
| DEPLOYwith a reminder set for December Gemini 3.8 Flash. $0.75 per million input and $3.75 per million output until December 31, 2026, then $1.50 and $7.50 from January 1, 2027. Google |
| WAITuntil you have read your contract Muse Spark 1.3. Two prices for the same model and the same 1M context window: $0.10 per million input on the contributor tier, where your inputs improve Meta's products, and $1.25 on the standard tier, where they do not. Meta |
| DEPLOYif you needed the license K2 Horizon. Open weights under Apache 2.0: 375B parameters, 23B active per token, a 512K context window, and a GPQA Diamond score of 87.3 in the model card's own table. IFM |
The pattern
Every row in this week's Deploy Log is a price, and no two are priced the same way. Anthropic held its headline rate and cut cache reads to $0.25 per million. Google set a date: $0.75 per million until December 31, 2026, then double from January 1. Meta charges $0.10 per million if your inputs improve its products and $1.25 if they do not. IFM removed the per token price and handed over the weights under Apache 2.0.
So the question The Deploy Log asks is not what a model costs, it is what it costs you and for how long: a rate that can be scheduled upward, a discount paid for with your data, or a license that moves the cost onto your own hardware. The lead is the one item this week that is not about price, and it is the one where a vendor rated its own model Critical and shipped it.