OpenAI makes GPT-6 Astra 50% faster
OpenAI just boosted the default inference speed for GPT-6 Astra and GPT-6.1 Sol by around 50%, pushing output from 30 to 50 tokens per second. The rollout applies across all subscription tiers and partner apps like Devin and OpenCode, kicking off a promised 28-day streak of daily shipping.
OpenAI kicked off a 28-day shipping sprint by boosting default inference speeds for GPT-6 Astra and GPT-6.1 Sol by roughly 50%, bumping output from 30 to 50 tokens per second.
Free speed boost for Astra and Sol
The rollout takes effect automatically across all subscription tiers within two hours, requiring no action from users.
Third-party products using Sign in With ChatGPT also get the speed boost immediately, including OpenCode, Pi, Amp, and Devin. OpenAI noted these models feature an optimized tokenizer as well, meaning they burn fewer tokens to finish the same work.
This marks Day 1 of a month-long sprint where OpenAI plans to ship a clear improvement or a full reset for work and coding users every single day.
Why it matters
For developers leaning on tools like Devin or Amp, faster token generation paired with a leaner tokenizer directly cuts waiting time and trims API usage.
Pushing this upgrade seamlessly to third-party tools via ChatGPT logins is a smart ecosystem play that reinforces OpenAI as the default backend for AI-assisted work. That said, hitting 50 tokens per second feels more like an incremental quality-of-life tune-up than a breakthrough, so the real test will be whether OpenAI can keep shipping meaningful upgrades for four straight weeks.
Sources (1)
- x.com(opens in a new tab)Originalx.com