Anthropic is preparing to release a new AI model that better understands user intent, The Information reported on Oct. 19.

The model is said to have stronger reasoning and intuition than Anthropic’s current models, the report said. A source who used the model said it is particularly good at picking up what users really mean even when they do not clearly state their goals, The Information reported.

For example, if a user asks it to fix a bug, the model identifies underlying causes and asks whether the user wants it to fix related problems at the same time. It also offers unconventional solutions rather than repeating obvious answers, The Information reported.

It is not yet known whether the model belongs to Anthropic’s top-tier "Fable" line or the next line down, "Opus".

Google is also developing a new model under the codename "Gemini 4 Carbon". It appears to be one level above its recently released "Gemini 4 Argon".

Argon did not outperform the top models from Anthropic and OpenAI, but developers rate Gemini models as much better for tasks requiring visual understanding, such as video analysis or checking whether a website works, The Information reported.

OpenAI is also expected to release a new model within weeks. It is likely to be positioned in the mid-tier "Terra" line. The model is expected to deliver higher performance and be cheaper than "GPT-6.1 Sol". The strategy is aimed at taking on Anthropic on price, The Information reported.

OpenAI, Anthropic and Google have released more than 10 new models over the past month, The Information reported. It added that more companies and developers are not automatically choosing the highest-performance models.