AI21 LabsTranslation site

What is AI21 Labs

AI21 Labs is an Artificial Intelligence startup founded on November 1, 2017 based in Tel Aviv, Israel. The company was founded by Amnon Shashua, Yoav Shoham, and Ori Goshen, three founders with team members having deep backgrounds in Artificial Intelligence and Computer Science.Focusing on Natural Language Processing (NLP) and generative AI technologies, AI21 Labs is dedicated to developing advanced AI models and delivering enterprise-grade AI solutions and consumer applications. Its main products include AI21 Studio developer platform, Wordtune writing assistant, Jurassic series of large language models, and Maestro enterprise AI platform. ai21 Labs has also made remarkable achievements in financing, obtaining investments from Google, Nvidia, and other well-known tech companies, which further consolidates its leading position in the AI field.

AI21 Labs History

1. Early development (2017-2020)
   • Company Formation and Team BuildingThe core team has established the company's technical foundation in the NLP field by virtue of its profound technical background and industry experience.
   • Wordtune Released: In 2020, Wordtune, an intelligent writing assistant, was launched and the number of users quickly surpassed 10 million, accumulating a large user base for the company.
2. Product Launch and Financing (2021-2023)
   • Jurassic-1 Jumbo Released: The launch of a large-scale language model with a parameter size of 178 billion in 2021, with performance close to GPT-3, marks a major technological breakthrough for the company.
   • Financing History::
     • Series C Financing: Closed a $155 million funding round in August 2023, valuing the company at $1.4 billion, with investors including Google, NVIDIA, and others.
     • C+ Series Financing: $208 million oversubscription closed in November 2023, increasing total financing to $336 million.
3. Technological breakthroughs and market expansion (2024-2025)
   • Maestro Platform ReleaseLaunched an enterprise-grade AI platform in 2025, which reduces the "illusion" rate by 501 TP4T and increases the inference accuracy to over 951 TP4T through the "hybrid expert" architecture and state-space modeling, applicable to high accuracy scenarios such as law, medicine, and finance. It is suitable for legal, medical, financial and other high-precision scenarios.
   • Series D Financing: 2025 Closed a $300 million financing round in May, led by Google and NVIDIA, for a total of $636 million, further solidifying the company's leadership position in AI.

AI21 Labs Products & Services

1. AI21 Studio
   • specificities: An NLP-as-a-Service developer platform that provides pre-trained and customized language models to support organizations in building text-based applications and services.
   • dominance: Flexibility and support for multilingual tasks for scenarios such as content generation and dialog systems.
   • application scenario: enterprise-level application development, academic research, content creation, and more.
2. Wordtune
   • specificities: AI-based writing and editing assistant with support for text rewriting, summarization, and grammar correction.
   • dominance: User-friendly, easy to operate, and supports multiple languages.
   • application scenario: business writing, academic writing, creative writing, etc.
   • Customer Feedback: With over 10 million users, it has been widely acclaimed, especially for improving the efficiency and quality of writing.
3. Jurassic Series Models
   • specificities: Includes Jurassic-1 and Jurassic-2 with a parameter size of 178 billion and support for multilingual tasks.
   • dominance: Powerful performance for complex text generation and comprehension tasks.
   • application scenario: content generation, dialog systems, machine translation, etc.
4. Maestro Platform
   • specificities: Enterprise AI platform that optimizes the accuracy and controllability of large models in real-world business through "hybrid expert" architecture and state-space models.
   • dominanceThe "hallucination" rate is reduced by 50%, and the inference accuracy is increased to over 95%, which is suitable for high-precision scenarios.
   • application scenario: Data-sensitive industries such as legal, medical, and financial.