Multimodal AI Topic - AI Updates

Abstract divide between European institutions and a multimodal artificial intelligence network

Meta will keep its forthcoming multimodal Llama model out of the EU, deepening concerns that regulatory uncertainty is creating a regional AI gap.

Real-time AI system processing voice images video and text simultaneously

View Post

Foundation Models

SenseTime Launches SenseNova 5.5 for Real-Time Multimodal AI

SenseTime’s SenseNova 5.5 combines real-time multimodal interaction with faster inference, low-cost edge deployment and enterprise migration support.

Five connected AI research systems representing images text music speech and diversity

View Post

Models & Research

Meta Releases Five AI Projects Across Images, Music and Speech

Meta’s FAIR releases combine multimodal generation, efficient model training, controllable music, speech watermarking and research on image diversity.

Multimodal AI assistant processing text video audio and code in real time

View Post

Foundation Models

Google Expands Gemini With Flash, Longer Context and Astra

Google’s Gemini update combines a faster model, two-million-token context, open vision models and a prototype universal assistant.

Person interacting naturally with an AI system through voice vision and text

View Post

Foundation Models

GPT-4o Unifies Voice, Vision and Text in One Model

OpenAI’s natively multimodal model reduces conversational delay while improving multilingual, audio and visual interaction.

Images charts and documents flowing into a compact multimodal neural model

View Post

Models & Research

Hugging Face’s Idefics2 Opens Vision-Language AI to Developers

The open 8B-parameter Idefics2 model combines improved OCR, native-resolution image handling and immediate Transformers support for multimodal applications.

Multimodal developer API analysing an image interface sketch and long document

View Post

Developer Platforms

GPT-4 Turbo With Vision API Moves Into General Availability

OpenAI’s generally available multimodal API combines image analysis, a 128K context window, JSON output and function calling for production applications.