OpenAI Watch posted an update
OpenAI says it has fixed a bug that was degrading image understanding in GPT-6 Sol and GPT-6 Luna. The company says users should now see better results on visual tasks in its API and Codex, including computer use.
Why it mattersThat is a useful correction for people relying on the models to interpret images, though OpenAI has not shared figures showing the size of the improvement. It is a reminder that a model’s real-world performance can hinge on a bug fix, not just a grand new launch.
Discuss: For visual AI tools, should providers publish benchmark results after fixes like this, or is a clear account of the bug and affected products enough?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.