Glean Claims 70% Lower Token Use Than Claude Cowork in Enterprise Task Test

Glean says its AI assistant used 70% fewer tokens than Anthropic’s Claude Cowork across more than 180 enterprise tasks, sharpening the startup’s pitch that better context and model selection can reduce the cost of workplace AI. The company also reported that its answers were preferred 78% of the time and that average token cost per task fell to $0.58, compared with $2.98 for Claude Cowork. Glean’s benchmark summary