Voyage AI

📢 We’re excited to announce our first multimodal embedding model: voyage-multimodal-3! voyage-multimodal-3 is capable of vectorizing interleaved texts + images, capturing key textual and visual features from screenshots of PDFs, slides, tables, figures, etc. voyage-multimodal-3’s unique architecture allows it to improve retrieval accuracy over the next best-performing model by 19.63% when evaluated across 3 multimodal retrieval tasks (20 total datasets). Most multimodal embedding models (e.g. Amazon Titan Multimodal G1, Cohere multimodal v3, etc) mimic OpenAI’s CLIP and use separate networks for images & text. In contrast, voyage-multimodal-3 mimics the architecture of modern vision-language models, processing both text and visuals within the same transformer encoder. The architecture of CLIP also prevents it from being usable in mixed-modality searches, as text and image vectors often align with irrelevant items of the same modality. With voyage-multimodal-3, there is no longer a need for screen parsing models, layout analysis, or any other complex text extraction pipelines. Simply take a screenshot of your document, prepend or append extra text (such as meta information), and convert the interleaved data into a single, unified vector. Check out our sample notebook to get started: https://lnkd.in/gj7JGuSV. For more details, check out our latest blog post: https://lnkd.in/gPbfkD5A Start building with @VoyageAI today - the first 200M tokens are on us! We’d also love to support academic retrieval research and benchmarking. Please email us at [email protected] for more free tokens.

We’re excited to share that #DatabricksVentures has invested in the Series A funding round of Voyage AI! Many enterprises are leveraging Compound AI Systems, which include RAG architecture, to address the GenAI accuracy gap and get their applications into production. A top-quality embedding model is a cornerstone of an accurate RAG system, and Voyage AI offers some of the world’s best embedding models available. We will also soon offer Voyage AI’s latest generation of embedding and rerank models natively within the Mosaic AI Model Serving solution, continuing our goal of helping customers build data intelligence and high-quality production AI systems. Stay tuned for more updates! https://dbricks.co/47WX52y

We're excited to support Voyage AI’s models in Cortex AI with the improved and advanced capabilities it will bring to our customers including multilingual capabilities and longer context windows. Read more in VentureBeat on how we're supporting and integrating Voyage AI into our Cortex AI services, and making AI more useful to enterprise users.

Thrilled to share that we've closed $28M in funding, led by CRV, with continued support from Wing Venture Capital and Conviction. Also excited to onboard strategic partners Snowflake and Databricks! We are on a mission to supercharge Search and Retrieval for Unstructured Data by building best-in-class embedding models and rerankers. We offer general-purpose, domain-specific, and customized embedding models + rerankers, with multimodal & more to come. Check out the latest generation: voyage-3 and voyage-3-lite: https://lnkd.in/g_5jpn_6 rerank-2 and rerank-2-lite: https://lnkd.in/gCDDbeAj Start building with Voyage AI today — Voyage models are now available via: 🔹 Voyage API 🔹 AWS Marketplace 🔹 Azure Marketplace 🔹 Snowflake Cortex AI 🔹 Licensing and on-prem A massive thank you to all of the folks who have supported us on this journey! Pear VC, Tectonic Ventures, Mayfield Fund, Fusion Fund, Zachary DeWitt, Sarah Guo, Pranav R., Christopher Manning, Fei-Fei Li, Michele Catasta, Amjad Masad, Eric S. Yuan, Tristan Handy, Ajay Singh, Mohit Aron, George Deglin, Dan Belcher, Erik Bernhardsson, @tanujt, Zheng Bu, Dylan Pearce, Bryan Subijano, Brian Zhan, Murat Bicer. https://lnkd.in/g7tW6AfU

📢 Announcing a new generation of natively multilingual Voyage rerankers: rerank-2 and rerank-2-lite! Adding rerank-2 and rerank-2-lite on top of @OpenAI’s latest embedding model (v3 large) improves accuracy by 13.89% and 11.86%, 2.3x and 1.7x the improvement attained by the latest @cohere reranker (English v3), respectively. Voyage rerankers work exceptionally well with Voyage embedding models - in the same benchmark, rerank-2 is the only reranker to improve atop voyage-multilingual-2. You can get a full list of the embedding models we provide at https://lnkd.in/gb6a3i2i For more details, check out our latest blog post: https://lnkd.in/g78DP787 Start building with @VoyageAI today - the first 200M tokens are on us! We’d also love to support academic retrieval research and benchmarking. Please email us at [email protected] for more free tokens.

📢 Announcing a new generation of Voyage embedding models: voyage-3 and voyage-3-lite! When compared with OpenAI's v3 large: voyage-3: + 7.5% accuracy, 2.2× cheaper, 3× smaller embedding dimension, 4x longer context voyage-3-lite: + 3.8% accuracy, 6× cheaper, 6× smaller embedding dimension, 4x longer context They are also natively multilingual. More in our latest blog post: https://lnkd.in/g_5jpn_6. Start building with Voyage AI today - the first 200M tokens are on us! We’d also love to support academic retrieval research and benchmarking. Please email us at [email protected] for more free tokens.

SPONSORED: Building GenAI is complicated. DataStax unveiled new integrations and solutions that make it easier for developers to get their GenAI apps to production.

🌍📢 Launching our multilingual embeddings, voyage-multilingual-2! 👑 Average 5.6% gain on evaluated languages, including French, German, Japanese, Spanish, and Korean 📚 32K context length 🛒 On AWS Marketplace Check us out! 👉🏼 The first 50M tokens are on us. More details in our blog post: https://lnkd.in/gVviqtF9 We support academic retrieval research and benchmarking! Please email us at [email protected] for more 🆓 tokens. #RAG #LLM

🆕 📢 Launching Voyage AI’s new embedding model for finance retrieval and RAG: voyage-finance-2! 1. ✨ Superior finance retrieval quality with an average of 7% gain over OpenAI and 12% over Cohere 2. 📚 32K context length 3. 🛒 On AWS Marketplace #RAG #LLM More detail in the blog post: https://lnkd.in/gwN4V_pR Check us out 👉🏼 The first 50M tokens are on us! We’d also love to support academic retrieval research and benchmarking. Please write us at [email protected] for more free tokens.

🆕📢 We are thrilled to launch rerank-1, our best general-purpose and multilingual reranker! It refines the ranking of your search results with cross-encoder transformers. 🔹 Outperforms Cohere’s english-v3 on the English dataset. 🔹 Outperforms multilingual-v3 on multilingual datasets. 🔹 8K context-length. More in blog post 📖: https://lnkd.in/gGQJGgaQ