Latest AI Updates: Video analysis by Google Gemini, video generation by Midjourney, and Bing’s video AI, among others.

In mid-June 2025, major tech companies such as Google, Microsoft, and Midjourney each announced significant upgrades to their AI services. These updates have greatly expanded the use cases for generative AI, advancing video creation and interactive information processing experiences.

1. Enhancement Update to Google Gemini Pro

Google announced the preview release of Gemini 2.5 Pro, with significantly improved AI reasoning and code generation capabilities. In particular, its ability to handle mathematical reasoning and complex programming tasks has been strengthened, making it a powerful tool for developers and researchers. A newly introduced feature called “Thinking Budget” allows users to finely tune the AI’s reasoning resources.

Additionally, the new “Deep Think” mode enables deeper analysis and idea development through parallel processing. These upgrades are gradually being made available through Google AI Studio and Vertex AI, and now also allow for more natural conversational expression and consistent long-form output.

2. Bing Video Creator: Microsoft’s AI Video Generation Tool

Microsoft has officially announced the release of Bing Video Creator, which allows users to generate short videos directly from text prompts.

This tool is based on OpenAI’s video generation technology and can instantly create short animations or explanatory videos from simple instructions. It is expected to be widely used in education, marketing, and social media content creation.

3. Midjourney Officially Releases Image-to-Video Feature

Midjourney has introduced its proprietary image-to-video generation model, making a notable impact in the video-generation AI field. Its main features are:

• Output Specs: Generates four 5-second videos simultaneously in 480p resolution. Videos can be extended infinitely. Supports custom aspect ratios.
• Cost: 1 video = the same cost as one image upscale. Plans starting at $60/month allow unlimited use.
• Mode Selection:
  • Low Motion for static scenes
  • High Motion for dynamic scenes
• Operation Modes:
  • Auto: Automatically extracts prompts from the image
  • Manual: Users input their own prompts
• Advantages: Works exceptionally well with Midjourney’s signature abstract/artistic style. Also compatible with older versions (v1~) and the Niji model.

4. Video Upload & Questioning Now Available in Gemini App

The Google Gemini app has finally added a local video analysis feature. Users can now upload video files from their smartphones to Gemini and ask questions such as:

• Checking the content of a specific scene
• Generating a summary of the video
• Extracting specific objects or lines from the footage

This feature is currently being rolled out to a limited number of users on the Android and iOS Beta versions of the Gemini app. Integration into the web version is planned for a future release.

Summary

• Google Gemini Pro Enhanced: The Gemini Pro model has been upgraded for faster, more accurate document comprehension, coding, and reasoning.
• Microsoft Bing Video Creator Released: Microsoft officially launches “Bing Video Creator,” a tool to generate short videos from text.
• Midjourney’s Image-to-Video Feature Released: Midjourney unveils its own model to generate 5-second 480p videos, excelling in artistic expression.
• Gemini App Supports Video Analysis: Google Gemini now allows users to upload and analyze local videos, enabling content-related questions and summaries.