The Information: Anthropic preps new model with sharper intent reading
Anthropic's upcoming model reportedly reasons better than its current ones, while Google tests Gemini 4 Carbon and OpenAI plans a Terra-line release.
Anthropic is preparing to launch a new AI model that reads user intent better than its current models, according to a report by The Information on October 19, 2026. The outlet reported that the upcoming model also shows stronger reasoning and intuition than Anthropic's existing models.
A source who tried the model told The Information that it is especially good at inferring a user's underlying goal even when that goal is not stated clearly. It is not yet known whether the new model belongs to the Fable line, Anthropic's top model tier, or the Opus line.
Google is also developing a new model under the codename Gemini 4 Carbon, apparently one tier above the recently released Gemini 4 Argon. OpenAI, meanwhile, is expected to release a new model within weeks, likely in its mid-tier Terra line, with higher performance and a lower price than GPT-6.1 Sol.
Developers weigh capability against fit
According to The Information, OpenAI, Anthropic and Google released more than 10 new models in the past month. Gemini 4 Argon did not outperform the top models from Anthropic and OpenAI, but developers rate Gemini models higher for tasks that require visual understanding, such as video analysis and checks on website operation.
The outlet added that a growing number of companies and developers are no longer automatically choosing the highest-performing models.
Quick answers
What is Anthropic's new AI model called?
It has not been reported. It is also not yet known whether the model belongs to the Fable line or the Opus line.
What is Google's Gemini 4 Carbon?
A model Google is developing under that codename, apparently one tier above the recently released Gemini 4 Argon.
When is OpenAI's next model expected?
Within weeks, likely in its mid-tier Terra line, according to The Information.