Palmyra X5, developed for efficiently powering multi-step agents, available exclusively via Writer and Amazon Bedrock as a fully managed model.
Amazon Web Services (AWS) has announced the availability of Palmyra X5—a new, state-of-the-art, adaptive reasoning model with a 1 million token context window—in Amazon Bedrock. Released by Writer, a leader in enterprise generative AI, Palmyra X5 is one of the first with a context window of this size to be available in Amazon Bedrock. Optimized for speed and cost efficiency, it empowers customers to build advanced, multi-step AI agents and transform how they reason over vast amounts of enterprise data with precision. AWS is the first and currently only cloud provider to deliver fully managed, serverless models from Writer, including both the new Palmyra X5 and Palmyra X4, with additional models available soon.
As generative AI continues to accelerate, customers need a wide choice of models to choose the ones best suited for their needs. The addition of Writer to Amazon Bedrock expands the already broadest selection of fully managed models available from leading AI companies for customers to easily and securely build and scale generative AI applications that transform and help them innovate across their businesses.
Palmyra X5 is one of the first 1M context window models available in Amazon Bedrock, further expanding the options available to AWS customers. (A context window is the amount of information a model can process and “remember” per input or request. It’s measured in tokens—the smallest units of text a model can handle—and can be thought of as the model's "short-term memory"). With a context window of this size, Palmyra can accurately handle 1500 pages of content, the equivalent of ~6 books. It is also one of the industry's first commercially available adaptive reasoning models. The model combines the capabilities of an advanced large language model (LLM) with expanded memory and processing capabilities. Enterprises can now tackle long-range tasks—financial reporting, legal-contract analysis, medical-record synthesis, customer-feedback mining, and more—without breaking the budget.
Beyond its reasoning capabilities, Palmyra X5 offers the ability to power agents that can interact with systems, advanced code generation and deployment, and multilingual support across more than 30 languages.
If Palmyra X5 were a person, they would have the superhuman ability to read 1 million words in 22 seconds, and immediately generate actionable insights. They're the team member who not only remembers everything in the 200-page strategy document, but also understands how it relates to yesterday's client meeting and last quarter's financial results. When faced with complex problems, they can methodically work through solutions step by step, explaining their thought process clearly along the way, whether they're helping analyze large volumes of customer feedback to identify common themes or troubleshooting a technical issue.
"We chose AWS as the first major cloud provider to deliver fully managed models from Writer because of its unmatched security and our shared vision of transforming how enterprises leverage AI to drive innovation and growth,” said Waseem AlShikh, Writer CTO and Co-founder. “Palmyra X5 is Writer's most advanced model yet, delivering the ability to process enormous amounts of enterprise data at speeds that are critical for scaling multi-agent systems. With Amazon Bedrock, we can bring these powerful capabilities into the hands of even more enterprises globally and empower our customers to deploy in Bedrock's secure, scalable environment."
"Building on our strong strategic collaboration with Writer, AWS is excited to offer Writer’s family of Palmyra models through Bedrock, enabling a new era of agentic innovation for organizations around the world,” said Atul Deo, Director of Amazon Bedrock at AWS. “Palmyra X5 offers impressive performance over long context inputs and enterprise-grade reliability and speed. Seamless access to Writer’s Palmyra X5 will enable developers and enterprises to build and scale AI agents and transform how they reason over vast amounts of enterprise data—leveraging the security, scalability, and performance of AWS."
Palmyra X5 is one of the most efficient large-context LLMs in the market, optimized for both speed and cost. Palmyra X5 can absorb an entire million-token prompt in ~22 seconds, and it can return individual function-calling turns in ~.3 seconds
Palmyra X5 demonstrates a best-in-class score-to-cost ratio in the latest Longbench v2 evaluation, with an impressive 53% average score. Enterprises get near top-tier accuracy while paying significantly less per million tokens—enabling a massive number of agentic and long-context tasks without breaking the budget.
It supports more than 30 languages, offering truly multilingual capabilities for global enterprises.
It’s priced at $.60 per 1 million input tokens and $6 per 1 million output tokens, making it one of the most cost-efficient large context LLMs in the market.
Palmyra X5 is also one of the top ranked models on the BigCodeBench (Full, Instruct) evaluation with a score of 48.7, showcasing its ability to tackle practical and challenging programming tasks.
While generative AI is changing how we create, analyze, and interact with information, agentic AI will fundamentally reshape the nature of work itself. This next frontier of AI moves beyond content creation and insight generation to using AI agents that can independently plan, execute, and adapt complex sequences of actions. With Palmyra X5 available in Amazon Bedrock, AWS customers can use Writer's model to securely and privately build and scale AI agents without managing underlying infrastructure. What's particularly exciting for enterprises across industries is that Palmyra X5 allows them to build and deploy more sophisticated AI agents that require larger amounts of data, including access to other agents, LLMs, and tools in external systems.
Writer ships precise, fully owned models that never undergo post-training quantization or distillation, so the behavior you validate today is the behavior you’ll see tomorrow. Palmyra X5 builds on that promise with strict backward compatibility to spare teams the pain of re-tuning flows, a published enterprise roadmap that customers can influence, and latency-optimized inference that makes LLM interactions and retrieval-augmented generation (RAG) feel instantaneous even at million-token scale.
Writer announced that it is now setting 1 million tokens as the minimum context window size for all LLMs it releases in the future, thanks to a novel transformer design (an architecture that can review input data simultaneously rather than sequentially) and hybrid attention mechanisms (allowing for multiple ways to focus on information at once for both efficiency and effectiveness). This means enterprises can plan their AI strategies around consistently expanding capabilities rather than being limited by context windows.
Visit the AWS News blog for more detailed information on Palmyra X5, including how to deploy the model in Amazon Bedrock and potential use cases, and check out the Writer in Amazon Bedrock product page.
Posted by :
Farih Khafiyyan
April 29th, 2025
Reach Us
We'd love to assist you. Fill out the form or drop us an email.
Visit Us
Drop by our office for a chat.
Gedung Pondok Indah Office Tower 3, 17th Floor, Jl. Sultan Iskandar Muda Kav V-TA, Jakarta 12310, Indonesia