Alphabet (Google) vs StepFun

A named tab in Google Search — "our most powerful AI search, with more advanced reasoning and multimodality" — that answers a query conversationally over web sources rather than with a link list.

Standalone mobile app (iOS/Android) delivering personalized, AI-illustrated daily stories synthesized from a user's connected Google services (Gemini, Gmail, Calendar, Photos, YouTube, Search).

Google's assistant for chat, research, image generation and work across Google apps.

A source-grounded research tool that summarizes, answers questions from and generates audio overviews of documents a user uploads.

AI filmmaking application for creating cinematic clips and scenes, built on Veo (video), Imagen (image) and Gemini (prompting) — the application layer above the Veo model already recorded in models.json.

Agentic music co-producer, formerly ProducerAI and graduated from Google Labs, that chats with you about ideas, generates and edits songs, learns your style, and offers free unlimited downloads.

An AI-assisted video creation and editing application bundled with Google Workspace that generates video from text prompts, slides or screen recordings.

A Google Labs science tool (built with Gemini Notebook) that runs a comprehensive literature search, structures results in a data table and produces reports, slide decks and infographics.

AI-powered concepting board for exploring, expanding and refining visual/creative ideas.

Google Labs experiment; explicitly waitlisted at evaluation time.

Enterprise search and retrieval service on the Gemini Enterprise Agent Platform (formerly Vertex AI Search) that indexes an organisation's websites, documents and structured data and serves search, answers and recommendations to applications and agents.

A managed Google Cloud API that transcribes audio to text, supporting streaming and batch recognition across 85+ languages.

A managed Google Cloud API that synthesizes natural-sounding speech from text across 380+ voices and 75+ languages.

A managed Google Cloud API that translates text and documents across more than 100 language pairs using neural machine translation, custom models or a translation-specialized LLM.

A managed Google Cloud API for image and document analysis, including object, face and landmark detection, OCR text extraction, and explicit-content tagging.

A pay-per-page API for extracting, classifying and summarizing structured data and text from documents such as invoices, forms and IDs.

ML-based text understanding: sentiment analysis, entity recognition, syntax analysis and content classification.

Video AI service that analyzes video content for object/label/shot detection and other video-intelligence tasks.

An agentic research engine (built with AlphaEvolve and ERA) that generates and scores code variations against the user's optimisation metrics to discover models and algorithms.

A multi-agent research tool (built with Co-Scientist) that simulates the scientific method to identify knowledge gaps and propose testable research plans.

An autonomous coding agent that clones a GitHub repository, plans and writes code changes in a cloud VM and opens a pull request for review.

Enterprise-grade notebook environment (IAM-managed access, Google Cloud security/network, official support) explicitly distinguished on its own page from the free Colaboratory product.

An AI coding assistant integrated into IDEs, GitHub pull requests and a CLI, sold in free, Standard and Enterprise tiers.

A free web tool for prototyping prompts and testing Gemini models before moving to the Gemini API for production use.

Agentic software-development platform managing autonomous coding agents across an IDE, a graphical command center and a terminal-first CLI surface.

Managed JupyterLab notebook instances on Google Cloud, prepackaged with deep-learning frameworks and integrated with BigQuery, Cloud Storage and scheduling, for developing and training models end to end.

A developer API, distinct from the consumer Gemini app, for integrating Gemini and Gemma models into applications and billed per token.

StepFun's consumer AI assistant, available on the web and as Windows and macOS desktop clients, offering web search, knowledge-base Q&A, image creation, tool use and speech recognition.

A subscription service from the StepFun Open Platform that supplies a monthly Credit pool for calling StepFun's flagship models from third-party coding tools and agent clients such as Claude Code, Cursor, OpenClaw, Cline and Zed.

A one-click cloud-deployed AI agent offered inside the 阶跃AI assistant that interprets instructions and carries out tasks autonomously.

A browser-based workspace for building, testing and deploying AI applications on StepFun's models, with a playground, a showcase library of example apps and a personal asset library.

API access to StepFun's text, vision, speech and image models, with credit packages and enterprise plans.