OpenAI makes GPT-6 Astra and GPT-6.1 Sol about 50% faster on ChatGPT subscriptions
In 30 seconds
OpenAI has optimized the default speed of GPT-6 Astra and GPT-6.1 Sol so they respond about 50% faster. The improvement applies when they are used through a ChatGPT subscription, in OpenAI products and in partner products with Sign in with ChatGPT, such as OpenCode, Pi, Amp and Devin. No changes are needed: it is day one of a 28-day challenge with an improvement or a usage limit reset every day.
OpenAI has optimized the default speed of GPT-6 Astra and GPT-6.1 Sol so they respond about 50% faster. Thibault Sottiaux, who works on Codex (the OpenAI coding agent) and ChatGPT at OpenAI, announced it in a post on X on Monday, October 5.
The improvement applies when these models are used through a ChatGPT subscription, across all OpenAI products and in partner products that let people log in with their ChatGPT account (Sign in with ChatGPT), such as the coding assistants OpenCode, Pi, Amp and Devin. No changes are needed: according to Sottiaux, the new speed should be felt within two hours of the announcement.
According to OpenAI, GPT-6 Astra is its most intelligent model, and GPT-6.1 Sol, introduced on September 29, nearly matches Astra on agentic coding, computer use and professional work at one-fifth of Astra's standard prices.
It is day one of the challenge Sottiaux announced on Sunday: for 28 days, each day they will either ship a clear improvement for most people who use Codex or ChatGPT Work, the ChatGPT agent for more ambitious tasks, or fully reset usage limits.
Why it matters
People who code with these models through their ChatGPT subscription get faster answers without changing anything.
Official source: Thibault Sottiaux (OpenAI) on X
