Intel / post

OpenAI fixes image-understanding bug in GPT-6 Sol and Luna

Source: x.com
Date: Published Sep 27, 2026

Context

OpenAI says a bug had been degrading image understanding in GPT-6 Sol and GPT-6 Luna, and that the bug is now fixed. In a post from its developer account on September 27, the company said users should now see better results on visual tasks in the API and in Codex, including computer use. Computer use is the mode in which a model works a computer by looking at screenshots and deciding where to click or type, so it depends on the model reading images correctly.

The fix came five days after OpenAI launched GPT-6 Sol and Luna on September 22 with API prices 50% lower than GPT-5.6. The post does not say when the bug began, how much it lowered quality, or which visual tasks it hit hardest.

Terms in this piece · Glossary

eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.

Read the source x.com