Microsoft’s BitNet Leads Innovative AI Developments: Flex Pricing from OpenAI, Google’s Gemini Advances, and Major Mobile Partnerships

Microsoft’s launch of BitNet, a one-bit open-source LLM optimized for standard CPUs, could revolutionize AI accessibility, while OpenAI’s Flex Processing offers cost-effective solutions for low-priority tasks. Google debuts Gemini 2.5 Flash with controllable reasoning, enabling users to adjust how the AI “thinks.” Meanwhile, Perplexity AI becomes Motorola’s default assistant and seeks expansion with Samsung. Lastly, xAI enhances its Grok chatbot with memory capabilities for more personalized interactions, entering the competitive field with OpenAI and Google’s AI assistants.

Table of Contents

• Microsoft Unveils BitNet
• OpenAI Launches Flex Pricing
• Google Launches Gemini 2.5 Flash
• Perplexity AI Becomes Default Assistant
• xAI Enhances Grok with Memory Feature
• Closing Thoughts

Hello everyone!
This is Aurora from AI Tech News Today, and today we have some exciting developments in the world of artificial intelligence. It’s April 18th, 2025, and here’s what’s making waves in AI this week.

Microsoft Unveils BitNet

Introducing a Revolutionary AI Model

Microsoft researchers have introduced BitNet, an impressively efficient open-source one-bit large language model featuring:

• Two billion parameters
• Trained on a staggering four trillion tokens

Key Features:

• Designed to run on standard Central Processing Units (CPUs)
• Requires less memory than most models

Potential Impact:
This could be a game changer for tech developers!

OpenAI Launches Flex Pricing

A New Way to Access AI

OpenAI is shaking things up with the introduction of Flex Processing, a new pricing tier for its application programming interface (API).

• This model allows users to opt for slower processing speeds at significantly reduced costs.
• Currently in beta for their o3 and o4-mini models.

Who Benefits?
This option aims to cater to lower-priority tasks like data enrichment and model evaluations.

Google Launches Gemini 2.5 Flash

Faster AI with Thinking Control

Google has rolled out Gemini 2.5 Flash, a preview version of its latest AI offering.

• This model boasts both high performance and developer-friendly options.

New Feature:
Introducing controllable reasoning, which allows users to adjust the model’s “thinking” process before it generates responses.

Perplexity AI Becomes Default Assistant

Mobile Integration Milestone

In a significant move, Perplexity AI has secured a partnership with Motorola to become the default assistant on its devices.

• Early discussions are also underway with Samsung Electronics to expand Perplexity’s reach across their Android ecosystem.

xAI Enhances Grok with Memory Feature

Personalized AI Experience

The tech firm founded by Elon Musk, xAI, has upgraded its Grok chatbot to now include a memory feature.

• This enhancement enables Grok to remember past user interactions.

Why This Matters:
It allows Grok to provide more personalized responses, a key improvement as it competes with OpenAI’s ChatGPT and Google’s Gemini, both of which already offer memory functionalities.

Closing Thoughts

That’s a wrap on this week’s AI highlights! Remember, the landscape of artificial intelligence is always evolving, with new developments at every turn.

Stay Informed!

Keep tuning in to stay updated on the latest trends and developments in the AI world!