#1 Benchmarked Palmyra-Med-70b and Palmyra-Fin-70b now available as open models on NVIDIA API catalog, Baseten, and Hugging Face, empowering developers with powerful, highly accurate tools to build and deploy industry-specific AI applications.
Palmyra-Fin-70b becomes the only financial services model available as an NVIDIA NIM inference microservice and the first LLM that can pass the CFAⓇ Level III Exam.
Writer, the full-stack generative AI platform delivering transformative ROI for the world’s leading enterprises, today announced Palmyra-Med-70b and Palmyra-Fin-70b, two industry-leading specialized large language models (LLMs) specifically designed for the healthcare and financial services industries. Beginning today, these cutting-edge models are available as open models from NVIDIA, Baseten, and Hugging Face, as well as via the Writer Platform. Theybring unparalleled domain expertise and accuracy to medical and financial generative AI applications, surpassing the performance of comparable models with single (zero-shot) attempts.
“Domain-specific LLMs are revolutionizing the AI landscape by delivering unmatched precision,” said Writer CTO and Co-Founder Waseem Alshikh. “Building on our track record of bleeding-edge innovation, Writer is pioneering this movement by creating models with deep, sector-specific expertise, exceptionally well-suited for enterprises use cases. We’re committed to advancing the health and financial industries hand-in-hand with our customers, ensuring our generative AI solutions drive transformative outcomes for the world’s leading companies and the people they serve.”
Palmyra-Med-70b and Palmyra-Fin-70b join Writer’s family of top-benchmarked LLMs, setting a new standard for domain-specific models. Extensive testing confirmed the superiority of both models over larger, generalized models like GPT-4 and specialized models like Med-Palmyra-2 and comparable financial models. Their exceptional performance across various domain-specific benchmarks underscores their remarkable ability to handle complex industry-specific tasks with exceptional accuracy.
Palmyra-Fin: Empowering Financial AI
Palmyra-Fin-70b stands as a game-changer in financial generative AI applications. It is now the only financial services model offered as an NVIDIA NIM inference microservice and available from ai.nvidia.com.
Palmyra-Fin-70b shows unparalleled proficiency in financial forecasting, investment analysis, risk evaluation, fraud detection, and asset allocation strategy, and it can power these use cases with unprecedented accuracy. Leveraging a vast corpus of financial data and incorporating state-of-the-art fine tuning techniques, Palmyra-Fin possesses specialized knowledge that enables it to understand and analyze financial information with extraordinary precision. Financial institutions and businesses can harness its capabilities to gain actionable insights and make informed decisions.
In a long-fin-eval benchmark test, Palmyra-Fin outperformed popular models like Claude 3.5 Sonnet, GPT-4o, and Mixtral-8x7b — showcasing the model’s ability to analyze complex financial topics. It is also the first model capable of passing the CFAⓇ Level III exam, often branded as “the world’s hardest exam” and one of the highest distinctions in the investment management profession.
Combined with the Writer full-stack generative AI platform — which includes integrated graph-based RAG technology, AI guardrails, and a suite of developer tools — Palmyra LLMs have enabled leading financial institutions like Vanguard and Franklin Templeton and innovative fintech companies like Intuit to reinvent their workflows with generative AI.
Palmyra-Med: Transforming Healthcare AI
Writer’s platform is also used by world-leading Fortune 50 healthcare companies, as well as healthcare innovators like Vizient, CirrusMD, and Medisolv, to help improve patient outcomes with powerful generative AI applications that are infused with deep medical knowledge.
Palmyra-Med-70b, the latest iteration of Writer’s healthcare models, is now the most accurate model available on the market. Rigorous testing revealed an impressive average accuracy of 85.9% across all medical benchmarks in a zero-shot attempt, dramatically surpassing the runner-up, Med-PaLM-2, which achieved 84% accuracy after six attempts.
Palmyra-Med-70b demonstrates profound proficiency in meeting the specific demands of the healthcare field. It excels in analyzing and summarizing complex clinical notes, EHR data, and discharge summaries. Its exceptional ability to identify key medical concepts, such as diseases, symptoms, medications, procedures, and anatomical structures, from unstructured text significantly enhances clinical decision-making. It also aids in the development and understanding of clinical trial protocols, drug interaction summaries, medical document generation, and much more.
Availability
Palmyra-Med-70b and Palmyra-Fin-70b are now available via Writer’s API, no-code tools, and the Writer Framework. They are also accessible on the NVIDIA API catalog, Hugging Face, and Baseten, enabling seamless deployment locally or to private clouds. Writer is releasing both Palmyra-Med-70b and Palmyra-Fin-70b under an open model license to give researchers, non-profits, and non-commercial users access to test and deploy the state-of-the-art models.
To explore the capabilities of Palmyra-Med and Palmyra-Fin and their potential to transform AI applications in healthcare and financial services, visit our blog. For more information on licensing for commercial use cases, get in touch with the Writer sales team.
Explore AITechPark for the latest advancements in AI, IOT, Cybersecurity, AITech News, and insightful updates from industry experts!
