BREAKING : GPT-5.2 Instant, Thinking, and Pro are rolling out on ChatGPT. Free and Go users will receive it a bit later. Also, GPT-5.2 Pro (High) is SOTA for ARC-AGI-2, scoring 54.2% for $15.72/task! Big at coding and math
GENERATIVE AI
-

GPT-5.2 stats exclude Grok: Is there beef between companies?
By
–
GPT-5.2 comparison stats exclude Grok… I wonder if the two companies have a beef or something
-
Runway’s new world model revolutionizes robotics and virtual worlds
By
–
Runway just changed the game.
— Robert Scoble (@Scobleizer) 11 décembre 2025
New world model.
Deep impacts on robotics, virtual beings, AI films, virtual worlds, and user interfaces of the future.@3duaun you can tell the Holodeck deniers to pound sand. @kevinafischer and @okpasquale I can see how your virtual beings will… https://t.co/YiZjJNmuI0Runway just changed the game. New world model. Deep impacts on robotics, virtual beings, AI films, virtual worlds, and user interfaces of the future. @3duaun you can tell the Holodeck deniers to pound sand. @kevinafischer and @okpasquale I can see how your virtual beings will
-
RunwayML’s World Models and high-quality avatars are a huge deal
By
–
OK, I think everyone is missing just how big a deal @runwayml
's World Models, just announced, are, along with a raft of other new things, including new avatars that really are virtual beings that are very high quality. Hey @TrishaCode remember how for more than a year now I -

Google Releases Gemini Deep Research Agent
By
–
BREAKING : Google released Gemini Deep Research Agent on API, based on Gemini 3 Pro as a core! Gemini Deep Research achieves state-of-the-art 46.4% on the full Humanity’s Last Exam (HLE) set, 66.1% on DeepSearchQA and a high 59.2% on BrowseComp
-
AGI-level convenience with current best models available
By
–
Right? The goal is to give you AGI-level convenience with today’s best models.
-
Cohere Launches Rerank 4 Fast and Pro Models
By
–
Rerank 4 Fast and Rerank 4 Pro are available today on our platform, @aws SageMaker, and @microsoft Foundry. Find more details in our blog:
-

Rerank 4 Supports 100+ Languages for Advanced Retrieval
By
–
Users can deploy Rerank 4 in over 100 world languages, including state-of-the-art retrieval in 10 major business languages.
-

Rerank 4 Enables Domain-Specific AI Model Customization
By
–
Rerank 4 offers a set self-learning capability with automated adaptation to domain-specific workflows. This lets users customize the model for particular use cases without the need for annotated data – including content types, language and industry terminology.
