Semana de IA em Revisão 26.02.07 - Substack

OpenAI releases GPT-4o mini

OpenAI has released a smaller, cheaper model called GPT-4o mini. The model is priced at $0.15 per million input tokens and $0.60 per million output tokens, making it significantly cheaper than GPT-3.5 Turbo. The model is designed for high-volume, simple tasks and is available in the API and ChatGPT.

Google launches NotebookLM Plus

Google has launched NotebookLM Plus, a paid version of its AI-powered note-taking app. The new version includes features such as larger usage limits, expanded capabilities, and enterprise-grade data governance and security. NotebookLM Plus is available as an add-on subscription for Google Workspace customers.

Stability AI releases Stable Diffusion 3.5 Large

Stability AI has released Stable Diffusion 3.5 Large, a new text-to-image model that offers improved image quality and prompt adherence. The model is available under a non-commercial license and can be accessed through Stability AI's API and partners like Amazon Bedrock and Fireworks AI.

Apple previews new accessibility features

Apple has previewed several new accessibility features coming later this year, including eye tracking, music haptics, and vocal shortcuts. The features are designed to make Apple devices more accessible to people with disabilities.

Meta unveils Purple Llama 2

Meta has unveiled Purple Llama 2, an open-source platform for developing responsible generative AI models. The platform includes tools for evaluating and fine-tuning models, as well as a collection of models and datasets.

Amazon announces Rufus

Amazon has announced Rufus, a new AI-powered shopping assistant that will help customers find products and make purchase decisions. Rufus will be available in the Amazon Shopping app in the coming weeks.

Google releases Gemma 2 2B

Google has released Gemma 2 2B, a small language model designed for edge devices. The model is available under a non-commercial license and can be used for tasks such as text generation and summarization.

Microsoft announces Phi-3.5-mini

Microsoft has announced Phi-3.5-mini, a small language model designed for on-device use. The model is available under a non-commercial license and can be used for tasks such as text generation and summarization.

IBM releases Granite 3.0

IBM has released Granite 3.0, a new family of open-source AI models designed for enterprise use. The models are available under a non-commercial license and can be used for tasks such as text generation and summarization.

Adobe announces Firefly Image 3 Model

Adobe has announced Firefly Image 3 Model, a new text-to-image model that offers improved image quality and prompt adherence. The model is available under a non-commercial license and can be used for tasks such as image generation and editing.

Perplexity announces Enterprise Pro

Perplexity has announced Enterprise Pro, a new paid version of its AI-powered search engine. The new version includes features such as larger usage limits, expanded capabilities, and enterprise-grade data governance and security.

Google releases PaliGemma

Google has released PaliGemma, a new open-source vision-language model. The model is designed for tasks such as image captioning and visual question answering.

Microsoft announces Phi-3.5-vision-preview

Microsoft has announced Phi-3.5-vision-preview, a new vision-language model designed for tasks such as image captioning and visual question answering. The model is available under a non-commercial license and can be used for research purposes.

OpenAI announces GPT-4o fine-tuning

OpenAI has announced GPT-4o fine-tuning, a new feature that allows users to fine-tune the GPT-4o model for specific tasks. The feature is available to developers through the OpenAI API.

Google releases Gemma 2 9B

Google has released Gemma 2 9B, a larger version of its open-source language model. The model is available under a non-commercial license and can be used for tasks such as text generation and summarization.

Stability AI releases Stable Video 3D

Stability AI has released Stable Video 3D, a new model that can generate 3D videos from text prompts. The model is available under a non-commercial license and can be used for tasks such as video generation and editing.

Microsoft announces Phi-3-vision

Microsoft has announced Phi-3-vision, a new vision-language model designed for tasks such as image captioning and visual question answering. The model is available under a non-commercial license and can be used for research purposes.

Google releases Gemini 1.5 Pro

Google has released Gemini 1.5 Pro, a new version of its large language model. The model is designed for tasks such as text generation and summarization and is available through the Google AI Studio and API.

OpenAI announces GPT-4o API

OpenAI has announced the GPT-4o API, a new API that allows developers to access the GPT-4o model. The API is available to developers through the OpenAI platform.

Google releases Imagen 2

Google has released Imagen 2, a new text-to-image model that offers improved image quality and prompt adherence. The model is available under a non-commercial license and can be used for tasks such as image generation and editing.

Microsoft announces Phi-3

Microsoft has announced Phi-3, a new family of small language models designed for on-device use. The models are available under a non-commercial license and can be used for tasks such as text generation and summarization.

OpenAI announces GPT-4 Turbo

OpenAI has announced GPT-4 Turbo, a new version of its large language model. The model is designed for tasks such as text generation and summarization and is available through the OpenAI API.

Google releases Gemini 1.0

Google has released Gemini 1.0, a new version of its large language model. The model is designed for tasks such as text generation and summarization and is available through the Google AI Studio and API.

Microsoft announces Phi-3-mini

Microsoft has announced Phi-3-mini, a small language model designed for on-device use. The model is available under a non-commercial license and can be used for tasks such as text generation and summarization.

OpenAI announces GPT-4

OpenAI has announced GPT-4, a new version of its large language model. The model is designed for tasks such as text generation and summarization and is available through the OpenAI API.

Google releases Gemini Nano

Google has released Gemini Nano, a small language model designed for on-device use. The model is available under a non-commercial license and can be used for tasks such as text generation and summarization.

Microsoft announces Phi-3-small

Microsoft has announced Phi-3-small, a small language model designed for on-device use. The model