Comment by heaney-555

Why is this article based on a chart that has GPT-5.5 and Claude Opus 4.8 as the newest evaluated models? Those are very old now.

There have been massive improvements in computer use in GPT 5.6 and Claude 5.