Welcome to Get The Gist, where every weekday we share an easy-to-read summary of the latest and greatest developments in AI—news, innovations, and trends—all delivered in under 5 minutes! ⏱

In today’s edition:

• LG Launches New 'Exaone Deep' AI Model

• Google Launches FireSat to Detect Wildfires Faster

• Mistral AI Has Launched Mistral Small 3.1

• Google Docs Introduces AI-Powered Summary Tool

• And more AI news….

Top Developments

LG

LG Launches New 'Exaone Deep' AI Model

The Gist: LG AI Research has released its advanced reasoning AI model, Exaone Deep, as open source, positioning it against major AI players like OpenAI and Google. The model excels in logical reasoning and problem-solving, particularly in mathematics and the Korean language.

Key Details:

• Exaone Deep-32B performed on par with larger models despite having only 32 billion parameters.

• It achieved top scores in Korea’s College Scholastic Ability Test and advanced scientific reasoning benchmarks.

• LG also released smaller versions, including a lightweight model retaining 95% of the flagship’s performance.

• The AI model is expected to support professional and scientific applications, reinforcing LG’s global AI ambitions.

Google

Google Launches FireSat to Detect Wildfires Faster

The Gist: The first satellite in the FireSat constellation has launched, using AI to detect wildfires earlier and more accurately. This system aims to improve emergency response and reduce wildfire-driven carbon emissions.

Key Details:

• FireSat can identify fires as small as 5x5 meters within 20 minutes.

• Google Research, Muon Space, and other partners developed the project with funding from Google.org and the Moore Foundation.

• The AI-based system compares real-time satellite imagery with past data to detect wildfires.

• The full constellation will enhance wildfire response and contribute to climate change research.

Mistral

Mistral AI Has Launched Mistral Small 3.1

The Gist: Mistral AI has launched Mistral Small 3.1, an open-source model that outperforms competitors in text, multimodal understanding, and long-context tasks while maintaining speed and efficiency.

Key Details:

• The model supports an expanded 128k token context window and delivers 150 tokens per second.

• It surpasses proprietary models like GPT-4o Mini and Gemma 3 in reasoning, multilingual, and multimodal benchmarks.

• Designed for efficiency, it runs on a single RTX 4090 or a Mac with 32GB RAM.

• Available on Hugging Face and API platforms, with enterprise deployment options through Google Cloud and NVIDIA.

Quick Gist

Instagram is testing AI-generated comments to boost engagement by allowing users to post computer-generated remarks under their posts. (Read More)

YouTube Music expanded its AI-powered ‘Ask Music’ feature to iPhone and iPad, enabling personalized radio station creation from user prompts. (Read More)

Google may shift its TPU production partnership from Broadcom to MediaTek to optimize costs and reduce dependency amid growing AI computing demands. (Read More)

Google launched Gemini's Deep Research feature, an AI-powered research assistant that generates detailed reports with citations. (Read More)

Google's Gemini 2.0 Flash AI model can remove watermarks from images, sparking concerns over copyright infringement and its impact on artists. (Read More)

Google Docs introduced an AI-powered summary tool that generates and dynamically updates document summaries for better readability and efficiency. (Read More)

Adobe launched AI agents within its Experience Platform, including the Brand Concierge, to enhance customer interactions and compete with Salesforce. (Read More)

THAT’S IT FOR TODAY!