OpenAI ships GPT-6 Sol and Luna with input pricing from $0.10 per million tokens

reported by 3 outlets· 3 articles · confidence: high · first seen 2026-09-22 18:00 UTC

What this means for you

If you pay for a larger OpenAI model, Luna is the tier to re-price against: $0.10 per million input tokens, Sol at $2, both live in the API now. The number that decides whether to switch is the quality gap, and it is not in the sources — no scores, no harness, no evaluation date.

OpenAI released two models on 22 September, GPT-6 Sol and GPT-6 Luna, both trained with methods it compares to its GPT-6 Astra work. Luna costs $0.10 per million input tokens and $0.50 per million output tokens; Sol costs $2 and $10. MarkTechPost reports 50% lower pricing, without naming the baseline. Both are live in the API, ChatGPT Work and Codex, with improved prompt caching (reusing the unchanged part of a prompt across calls). OpenAI claims fewer mistakes; no benchmark scores, evaluation dates or harness are given.

Models in this story

Key facts

  • ·GPT-6 Luna is priced at $0.10 per million input tokens and $0.50 per million output tokens; GPT-6 Sol at $2 and $10. source
  • ·Both models are available now in the API, ChatGPT Work and Codex. source
  • ·Both ship with improved prompt caching for long-running agents. source
  • ·OpenAI says the two models are cut from the same cloth as Astra. source
  • ·The release is dated 22 September 2026. source

What the sources say

  • OpenAI News — The vendor's own announcement; names the two models and their capability-cost split, with no pricing.
  • TechCrunch AI — Frames the launch around OpenAI's own claims of lower cost and fewer errors, linking it to Astra.
  • MarkTechPost — Carries the token prices, availability across API, ChatGPT Work and Codex, and the prompt-caching change.

Sources

The original reporting. Follow these — they did the work.

Related stories

← the wire