AI News/AI News

OpenAI ships GPT-6.2 Astra and Sol with a 20% cheaper Astra API

The two models target professional and everyday work, with strong cost-performance results and some uneven third-party findings that content teams should weigh.

Key points

  • OpenAI released GPT-6.2 Astra and GPT-6.2 Sol to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, plus the API.
  • Astra API prices fall 20% to $8 input, $0.40 cached input and $40 output per million tokens; Sol keeps $2, $0.10 and $10.
  • Per Artificial Analysis, Astra ranks second on its Intelligence Index behind Anthropic's Claude Fable 5.2, but sits on the cost-performance frontier.
  • Outside testers found gaps: Vals AI ranked the models lower on legal tests, and Artificial Analysis saw flat coding scores and high hallucination rates.

Why it matters for marketers: Astra is cheaper per task and Sol is cheaper still, but results vary by workload, so teams should test on their own slides, documents and research tasks before switching.

OpenAI has released GPT-6.2 Astra, its frontier model for complex professional work, and GPT-6.2 Sol, a cheaper model for everyday tasks. Both are available now in ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu users, and through the API. Pro 200 subscribers get Ultrafast mode for the first time, starting with Sol.

OpenAI says Astra produces slides, spreadsheets and documents that follow templates and professional conventions more closely, down to font sizes in slide layouts. It also says early testers needed fewer edits on decks. Sol, the company says, comes within 0.5 points of Astra on Cognition's FrontierCode 2.0 at about one-fifth the cost. Most figures here are OpenAI's own.

Independent results are mixed. Artificial Analysis puts Astra second on its Intelligence Index, behind Claude Fable 5.2, while placing OpenAI's models on the cost-performance frontier. It also reports flat coding scores and hallucination rates of 59% for Astra and 55% for Sol on a deliberately hard factual test. Vals AI found the models ranked well below the top on legal benchmarks despite doubling their predecessors' scores.

For marketing and content teams, the practical question is cost per finished deliverable. Sol is the likelier default for routine drafting and agent workflows, while Astra suits template-heavy decks and research. Run both on your real briefs and templates, and check factual claims, since benchmark gains may not carry over to your work.

Sources

This brief is our own summary. Read the full reporting at the source.

You read about AI automation. We run it.

DocFluence, built by Z Media, writes, designs, films and publishes for your brand every day, in your voice, from one login.