A new architecture for GPT-6.1 Sol?
OpenAI halved its cached-input price compared with GPT-6 Sol. Our measurements show it also handles long prompts faster.
Cached input on GPT-6.1 Sol costs half that of GPT-6 Sol and long prompts run faster in Epoch's measurements. Agents with large repeated contexts get cheaper and quicker. Architecture change is speculation.
Terms in this piece · Glossary
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
