This media is not supported in your browser
VIEW IN TELEGRAM
BREAKING 🚨: Luma Labs is launching Ray 3 Modify, a new video generation model that allows users to modify existing videos by providing character reference images.
You can inject yourself everywhere now 👀
You can inject yourself everywhere now 👀
🔥9❤2
TestingCatalog AI News 🗞
OpenAI readies Codex Caribou model upgrade with GPT-5.2 base OpenAI is preparing to release an updated Codex model, internally called Caribou, likely based on GPT-5.2. This version aims to raise baseline coding capabilities and may influence industry benchmarks…
BREAKING 🚨: OpenAI is releasing GPT-5.2-Codex "Caribou" model today! An earlier code was updated with a new model name just recently.
🥰6❤4
TestingCatalog AI News 🗞
Google is likely about to release Gemma 4 today, as "Google's Gemma models family" collection got updates just recently. F5F5F5 👀
Not Gemma 4 but functiongemma-270m-it 👀
"FunctionGemma is intended to be fine-tuned for your specific function-calling task, including multi-turn use cases."
"FunctionGemma is intended to be fine-tuned for your specific function-calling task, including multi-turn use cases."
😭13👍4💩2
Seems like NotebookLM is about to get a big upgrade soon. A new model selector would be a good fit over there too.
Flash AI news ⚡️⚡️⚡️
Flash AI news ⚡️⚡️⚡️
❤7🔥5👍3
Meta AI now allows users to add themselves to AI images and videos, in a similar way to Cameos on Sora. Voice sample can be added as well.
Note that Meta is globally available while Sore is not.
Social AI race is here 👀
Note that Meta is globally available while Sore is not.
Social AI race is here 👀
🔥7❤2
Meta is developing a new image and video AI model “Mango”, along with a previously reported “Avocado” according to WSJ.
❤7🔥4
Mistral AI launches OCR 3 model for document parsing
Mistral AI released Mistral OCR 3, a model with a 74% win rate over its predecessor, offering accurate document parsing for forms, tables, and handwriting. It's available via API and AI Studio, with pricing starting at $2 per 1,000 pages.
🗞 #mistral
Mistral AI released Mistral OCR 3, a model with a 74% win rate over its predecessor, offering accurate document parsing for forms, tables, and handwriting. It's available via API and AI Studio, with pricing starting at $2 per 1,000 pages.
🗞 #mistral
TestingCatalog
Mistral AI launches OCR 3 model for document parsing
What's new? Mistral OCR 3 wins 74% over its prior version in document text extraction; it powers Document AI Playground on Mistral AI Studio with an API at $2 per 1,000 pages;
❤4👍2
Google debuts FunctionGemma for on-device AI function calling
Google has released FunctionGemma, a compact version of its Gemma 3 model optimized for function calling on edge devices. It supports a broad tech stack and enables local, privacy-focused AI agents that convert language into API actions.
🗞 #huggingface
Google has released FunctionGemma, a compact version of its Gemma 3 model optimized for function calling on edge devices. It supports a broad tech stack and enables local, privacy-focused AI agents that convert language into API actions.
🗞 #huggingface
TestingCatalog
Google debuts FunctionGemma for on-device function calling
What's new? FunctionGemma is a function calling model from Gemma 3 270M for API actions; it runs on NVIDIA Jetson Nano and mobile phones with a 256k JSON vocabulary;
❤2👍2