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:

• ElevenLabs Launches New Speech to Text Model

• OpenAI's Latest GPT Model to Launch on Android Soon

• Google Translate is Getting a Powerful AI Upgrade

• Microsoft Launches Phi-4-Multimodal and Phi-4-Mini

• And more AI news….

Top Developments

ElevenLabs

ElevenLabs Launches New Speech to Text Model

The Gist: ElevenLabs has launched Scribe, an advanced AI transcription model that delivers the most accurate speech-to-text results across 99 languages. It outperforms leading competitors in benchmarks and is designed for real-world audio applications.

Key Details

• Scribe achieves industry-leading accuracy, with the lowest word error rates in Italian (98.7%), English (96.7%), and many underserved languages.

• Features include word-level timestamps, speaker diarization, and non-speech event tagging for structured transcription.

• Available via API for developers and through the ElevenLabs dashboard for creators and businesses.

• A real-time low-latency version is coming soon for live transcription applications.

OpenAI

OpenAI's Latest GPT Model to Launch on Android Soon

The Gist: A notification in the ChatGPT Android app hints that OpenAI’s next model, GPT-4.5, could be launching soon. While details remain unclear, Pro users are expected to get early access when it officially rolls out.

Key Details

• Some Android users have spotted a "GPT-4.5 research preview" alert, though it isn’t functional yet.

• OpenAI has not officially announced a release date, but speculation suggests it could arrive soon.

• Updates to ChatGPT’s web interface are in testing, including a new file selector for better navigation.

• GPT-4.5 is expected to improve problem-solving with more human-like reasoning and step-by-step logic.

Google

Google Translate is Getting a Powerful AI Upgrade

The Gist: Google Translate is adding an AI-powered "Ask a follow-up" feature that lets users refine translations by adjusting tone, style, or clarity. While not widely available yet, a global rollout is expected soon.

Key Details

• The new feature allows users to tweak translations, making them more natural or fitting regional variants.

• Users can simplify text, adjust formality, or even make translations humorous.

• Google’s Gemini AI can work alongside Translate for even more fine-tuned results.

• The feature is currently in testing, with signs pointing to an imminent release.

Quick Gist

Microsoft launched Phi-4-multimodal and Phi-4-mini, advanced small language models available on Azure AI and other platforms, enhancing multimodal interactions and efficient processing capabilities for diverse applications (Read More).

French AI app Le Chat, developed by Mistral AI, achieved over 1 million downloads within two weeks of launch (Read More).

Microsoft expanded access to its Copilot AI tools, Think Deeper and Voice, now allowing unlimited use for all users after previously imposing limits (Read More).

The Bank of New York Mellon has partnered with OpenAI to enhance its AI platform Eliza and accelerate innovation in financial services through advanced AI capabilities (Read More).

Salesforce expanded its partnership with Google, integrating Google’s Gemini AI into Salesforce's Agentforce to enhance enterprise AI capabilities and streamline business operations (Read More).

THAT’S IT FOR TODAY!