Search AI Products and News

Recently, Google officially launched a new command-line tool - Gemini CLI. This tool is based on Google's self-developed Gemini2.5Pro AI model, aiming to provide developers with convenient AI Q&A and content generation services. With Gemini CLI, developers can directly call the power of AI in their own terminal interface, thereby improving programming efficiency and work convenience. One of the highlights of Gemini CLI is its support for up to 1 million tokens

Recently, Google accidentally published a blog post about the Gemini CLI, but the page is now returning a 404 error. The unpublished article revealed several key features of the Gemini CLI, sparking discussions in the developer community. The content indicates that the Gemini CLI is an open-source command-line interface (CLI) tool for Gemini2.5Pro, supporting a context window of up to 1 million tokens, suitable for handling complex code and large-scale data.

Xiyu Technology announced the launch of its latest R&D achievement - Hailuo Video Agent, an intelligent assistant for video creation aimed at simplifying the video creation process and enhancing creation efficiency, and has officially opened a beta version for user experience.

According to AIbase, AI code editor Cursor announced on June 17, 2025, a major upgrade for its Pro plan, officially removing the monthly 500 fast request limit and launching the long-anticipated 'unlimited use' mode. This move is seen as a milestone change by Cursor in responding to user demands and enhancing developer experience. AIbase pointed out that for a long time, the 500 fast request limit of the Cursor Pro plan has sparked extensive discussions among users. Many developers have reflected that in high-intensity coding environments

Alibaba Group announced the release of an upgraded version of its flagship AI model Qwen3, which has successfully been adapted to Apple's MLX architecture. This indicates that the launch of Apple Intelligence in the Chinese market may be imminent. According to Alibaba, the new Qwen3 version boasts stronger performance, supporting 119 languages and dialects, further enhancing user experience. In the previous iOS 18.4 update, Apple Intelligence had already started supporting Simplified Chinese, but its generation ...

At the recently held 'Force Primordial Power Conference', ByteDance’s Volcano Engine officially launched the DouBao large model 1.6 and video generation model Seedance 1.0pro. This release marks another leap forward in AI technology, particularly in multi-modal understanding and video generation, demonstrating strong market competitiveness. DouBao 1.6 has performed excellently in multiple international authoritative evaluations, especially excelling in complex reasoning and competition-level mathematics tests. Tan Dai, President of Volcano Engine, stated that this model not only supports multi-

At the FORCE Original Power Conference, Volcano Engine officially launched BeanPod Large Model 1.6 and its new video generation model Seedance 1.0 Pro, marking a comprehensive upgrade of AI cloud-native full-stack services. As a leading family of multimodal, high-value models, BeanPod Large Model 1.6 has undergone significant improvements in functionality and performance. ByteDance CEO Liang Ruobo stated at the launch event that BeanPod Large Model 1.6 adopts a unified pricing model, creating an input length-based pricing approach.

Figma officially announced the launch of the Beta version of its Dev Mode MCP (Model Context Protocol) server, marking further integration between design and development workflows. This service standardizes the protocol for directly transmitting Figma design data to AI coding tools, significantly improving the efficiency and accuracy of converting design into code. The Dev Mode MCP service is now officially available, with the Figma Dev Mode MCP server already in Beta.

OpenAI announced a major price cut of 80% for its flagship reasoning model o3 and simultaneously launched the high-performance o3-pro model. This move significantly reduces developer costs, triggering a new round of competition in the AI market. o3 Price Drop Significantly OpenAI has made significant adjustments to the pricing of the o3 model as follows: Input Tokens: Original price was $10 per million tokens, now reduced to $2 per million tokens. Output Tokens: Original price was $40 per million tokens, now reduced to $8 per million tokens. Cached Input Tokens: An additional 75% discount is provided.

Recently, OpenAI announced the launch of its latest o3-pro AI model, which is designed to provide more reliable and accurate responses for enterprises. As an advanced version of the o series inference models, o3-pro supports more software tool integrations and is particularly suitable for businesses and developers requiring high precision and accuracy. However, the response speed of o3-pro has slowed down compared to previous models. OpenAI stated that due to o3-pro's enhanced capabilities with richer tools, it needs extra time when handling complex problems.

Recently, OpenAI's ChatGPT has experienced some ups and downs, causing chaos among global users due to its downtime. The company quickly released the new-generation model — o3-pro. This upgraded model is more precise in handling complex problems and demonstrates significant advantages in scientific research, programming, education, and writing, becoming more intelligent and feature-rich. The o3-pro model is now available to Pro and Team users, replacing the original o1-pro, Enterp

The French AI lab Mistral recently announced the official release of its first inference model series -- Magistral. The series includes two versions: Magistral Small and Magistral Medium, aimed at enhancing logical reasoning capabilities in disciplines such as mathematics and physics. Mistral stated that the Magistral series models solve problems step by step to improve the consistency and reliability of results.

In the recent Fiction.Live benchmark test, Gemini2.5Pro performed excellently in understanding and reproducing complex stories and backgrounds, leading ahead of OpenAI's o3 model. This test goes far beyond traditional "needle-in-a-haystack" tasks, focusing on a model's ability to handle deep semantics and context-dependent information within large contexts. According to the test data, when the context window length reaches 192,000 tokens (approximately 144,000 words), the performance of the o3 model drops sharply, while