Open-source AI search plugin Evoke achieves a 66.7% recall rate, rivaling models with nearly 600 million parameters.

iconKuCoinFlash
Share
AI summary iconSummary
Intelligent Internet, founded by Emad Mostaque, co-founder of Stability AI, has open-sourced Evoke, a PostgreSQL search plugin powered by a 30.3 million-parameter AI model. Based on IBM’s Granite-Embedding-30M-Sparse, the model enables semantic and keyword search without requiring GPU support. Evoke achieved a 66.7% recall rate in English benchmarks, outperforming traditional search methods and nearly matching the performance of a 600 million-parameter model. The plugin is ideal for AI + crypto news applications, including enterprise document retrieval and AI agent knowledge bases. The code is available under the Apache-2.0 license. Interest rate news and AI advancements continue to shape the tech landscape.
ME AI message: Stability AI co-founder Emad Mostaque’s new company, Intelligent Internet, has open-sourced Evoke—a PostgreSQL database search plugin featuring an AI model with approximately 30.3 million parameters. After installation, developers can directly search documents and knowledge bases within the database, matching both keywords and semantically similar content. Evoke is built on IBM’s Granite-Embedding-30M-Sparse model, which generates a weighted set of relevant terms for each document, enabling retrieval even when the exact search term does not appear in the text. Keyword and semantic information share a single search index, eliminating the need for separate vector indexes. The model runs directly on standard CPUs without requiring GPUs or sending data to external AI services. Official English retrieval tests previously released showed that Evoke retrieves relevant documents in the top 100 results with a 66.7% success rate—higher than traditional keyword search’s 56.3% and approaching the 67.1% of a 600-million-parameter model. Both Evoke’s code and model are open-sourced under the Apache-2.0 license, making it suitable for enterprise internal document retrieval, AI agent knowledge bases, and similar use cases. (Source: BlockBeats)
Source:Show original
Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information. Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.