Asana Cuts Model Costs 76x in Browser Tests with GPT-6.1 Sol
OpenAI, Friday, October 9th, 2026
Using GPT-6 Astra in Codex, Asana optimized its StackAI browser agent on GPT-6.1 Sol to run 76x cheaper and 5x faster.
Asana's StackAI CTO used GPT-6 Astra in Codex to investigate the company's browser agent, test improvements and compare results across models.
In a 144-run study comparing GPT-6.1 Sol with three other frontier models, the optimized workflow on GPT-6.1 Sol ran 76x cheaper and 5x faster, with an average estimated model cost of $0.47 per run.
Work estimated to take one to two months by hand took about a week.