OpenAI introduced its newest frontier model, GPT‑6.2 Astra, alongside the streamlined GPT‑6.2 Sol, targeting higher performance and lower costs for demanding business applications.

OpenAI has unveiled its latest generation of language models, GPT‑6.2 Astra and the lighter GPT‑6.2 Sol, promising faster response times and reduced operating costs for enterprise‑level workloads.

What’s new in GPT‑6.2 Astra

Astra builds on the architecture of its predecessor with a larger parameter count and optimized tensor cores, delivering up to a 30% boost in inference speed for complex prompts. The model also introduces adaptive context windows that dynamically allocate resources based on the length and difficulty of the input, helping businesses handle extensive documents without sacrificing latency.

GPT‑6.2 Sol: a streamlined alternative

Designed for cost‑sensitive applications, Sol trims the model size while retaining core reasoning capabilities. It targets use cases such as routine email drafting, data entry automation, and basic analytics, where raw power is less critical than throughput and price efficiency.

Targeted professional use cases

Both models are positioned for sectors that demand high reliability, including legal document review, financial modeling, and customer support automation. OpenAI highlights that Astra’s enhanced reasoning can reduce the number of iterative prompts needed, while Sol’s lower compute footprint makes it viable for large‑scale deployment across call‑center bots.

  • Legal firms can automate contract clause extraction
  • Financial analysts can generate real‑time market summaries
  • Customer support teams can scale chat‑bot interactions with lower latency

OpenAI also rolled out new API pricing tiers aligned with the performance gains, allowing enterprises to select a plan that matches their workload intensity and budget constraints.

For a deeper look at the technical specifications and pricing details, see the OpenAI page on Momentum and explore related stories in the AI category.

“These models represent a meaningful step toward making AI both powerful and affordable for everyday business processes,” said an OpenAI spokesperson.

The Verge coverage of OpenAI's GPT‑6.2 Astra release provides additional context and developer commentary.

The Verge coverage of OpenAI's GPT‑6.2 Astra release