GPT-6 Astra: an automated AI Engineer you can hire for <$6 an hour
Source
latent.space
Date
Context
The claim that OpenAI's Astra costs under $6 an hour comes from a particular generation-rate figure. Latent Space's preview report pairs 33 tokenThe chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.Full definition → per second with an output price of $50 per million tokens. Sustaining that rate for an hour gives $5.94 in output charges, and the report notes the rate assumes its preview latency holds at general availability.
That figure is expressed in output charges for one stream of work. The report itself warns that Astra's ability to run work in parallel can drive spending well above $6 an hour. The useful distinction is between the cost of one stream of generated output and the total cost of completing a project.
Terms in this piece · Glossary
token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
