GROK PREND LA PAROLE : LE MODE VOIX EST LÀ Grok vient de lancer le Mode Voix, permettant aux utilisateurs de parler à l’IA et de l’entendre répondre – en temps réel. Imaginez Siri… mais plus sarcastique – et bien plus intelligent. Source : xAI
MULTIMODAL AI
-
Google Releases Gemini 2.0 Flash Update with Improved Image Generation
By
–
Google pushed an update to its Gemini 2.0 Flash image generation model
— Rowan Cheung (@rowancheung) 8 mai 2025
The release promises improved quality of generations with better text rendering and fewer content restrictionspic.twitter.com/SJP3jVKDQ1Google pushed an update to its Gemini 2.0 Flash image generation model The release promises improved quality of generations with better text rendering and fewer content restrictions
-
Meta Launches Perception Language Model for Advanced Video Analysis
By
–
Meta AI dropped Meta Perception Language Model, an open & reproducible vision-language AI for challenging visual tasks
— Rowan Cheung (@rowancheung) 8 mai 2025
It can watch videos and extract details like what a person is doing in the content and how they are doing itpic.twitter.com/JSbbgCukRIMeta AI dropped Meta Perception Language Model, an open & reproducible vision-language AI for challenging visual tasks It can watch videos and extract details like what a person is doing in the content and how they are doing it
-
Mistral Medium 3 matches top models at 8x lower cost
By
–
Mistral just made two big reveals:
— Rowan Cheung (@rowancheung) 8 mai 2025
—Medium 3, a multimodal AI matching or surpassing 3.7 Sonnet, GPT-4o, and Llama 4 Maverick across benchmarks with 8x lower costs
—Le Chat Enterprise with corporate tools like Google Drive, agent building, and morepic.twitter.com/DUVh8nhD5JMistral just made two big reveals: —Medium 3, a multimodal AI matching or surpassing 3.7 Sonnet, GPT-4o, and Llama 4 Maverick across benchmarks with 8x lower costs —Le Chat Enterprise with corporate tools like Google Drive, agent building, and more
-
Mistral Launches SOTA Multimodal AI at Lower Cost
By
–
AI NEWS: Mistral just launched a new multimodal AI promising SOTA performance at much lower cost Plus, more news from Meta, Anthropic, Microsoft, Google, Figma, and Netflix. Here's everything you need to know:
-
R³-VQA: Read the Room Video Social Reasoning
By
–
R³-VQA: “Read the Room” by Social Reasoning
Paper: https://
arxiv.org/pdf/2505.04147 -
R³-VQA Benchmark Tests AI Theory of Mind in Social Scenes
By
–
How well can AI read the room? R³-VQA is a new video QA benchmark pushing LVLMs to reason like humans in complex social scenes—tracking beliefs, emotions, intentions, and more. Turns out, even top models still struggle with consistent Theory of Mind.
-
PixVerse v4 video generation model passes TikTok dancer test
By
–
PixVerse v4 passes the "tiktok dancer on drone" test, albeit with a bit less coherency than Wan 2.1 (but it makes up for this with frame rate)https://t.co/QsZQhLtzfA https://t.co/RhiXYlG4st pic.twitter.com/ykId1gXRIE
— fofr (@fofrAI) 7 mai 2025PixVerse v4 passes the "tiktok dancer on drone" test, albeit with a bit less coherency than Wan 2.1 (but it makes up for this with frame rate) https://
replicate.com/p/6p4ypb98csrm
80cpnk59mwyn14
… -
Google advances Computer Use in AI Studio for agent deployment
By
–
Google keeps working on Computer Use in AI Studio. "Deploy, scale, and maintain virtual desktop instances for agents. This deploys a sample architecture to allow Gemini to use features such as browser access, code execution, and file system capabilities."
-
Try Ideogram 3.0 Models on Replicate Platform
By
–
Try out each of the models on Replicate: Ideogram 3.0 Turbo: https://
replicate.com/ideogram-ai/id
eogram-v3-turbo
…
Ideogram 3.0 Balanced: https://
replicate.com/ideogram-ai/id
eogram-v3-balanced
…
Ideogram 3.0 Quality: