GPT-5.6 sol Reported as Cheaper and More Token-Efficient

eyishazyer · x · 2026-07-15

Replying to @sama, a user noted that GPT-5.6 sol costs half as much and boasts roughly double the token efficiency compared to fable across many tasks, allowing the same work to be done at a lower cost.

The author added that under these metrics, they would also be willing to drop their own delivery price to a quarter.

Related event: OpenAI Launches GPT-5.6 Family: Three Tiers Roll Out with Focus on Sol Cost-Efficiency(10 posts)→

Original post →

More from Models

Models channel →