Writer has launched Palmyra X6, a new flagship AI model built for agentic and long-running enterprise tasks, alongside an upgraded orchestration system designed to reduce the cost of running AI workloads. The company says the combination can lower costs for basic tasks by as much as 50%.
Palmyra X6 is based on Z.ai’s open-source GLM-5.2 model and has been further adapted by Writer for enterprise deployment. It will sit alongside Writer’s existing models and third-party models available through platforms such as Microsoft Azure and Amazon Bedrock.
Writer CEO May Habib said the company is focusing less on benchmark gains and more on making AI deployments cheaper and more predictable for enterprise customers.
Writer Says the Harness Matters as Much as the Model
The company also updated its agentic harness, the orchestration layer that manages context, tools, prompts, and multi-step workflows around a model. Writer argues that optimizing this layer can have a larger impact on costs than switching between different foundation models.
That view is supported by a research paper from Writer published in July. Across 22 enterprise agent tasks and six models, researchers found that Writer’s harness reduced blended cost per task by 41%, from $0.21 to $0.12, while cutting median completion time by 44% and token usage by 38%.
Every model tested became cheaper to run, with cost reductions ranging from 33% to 61%. The study also found that quality per dollar increased by 82%, while completed tasks per million tokens rose from 54.9 to 92.0.
Palmyra X6 Targets Lower-Cost Agent Workflows
Writer says Palmyra X6 is designed for complex, multi-step tasks where excessive reasoning, tool calls, or repeated context can quickly drive up token consumption. Its updated harness attempts to reduce that overhead by managing context and execution more efficiently.
The company is not requiring customers to use Palmyra X6 exclusively. Writer remains model-agnostic, allowing businesses to combine its own models with external systems depending on the task.
Writer added Palmyra X6 and the updated agent experience to its product lineup in August, describing the model as its general-purpose flagship for agentic work and cost efficiency at scale.
The release comes as enterprise AI buyers pay closer attention to the total cost of running agents, particularly as workflows become longer and use more tools. Writer’s position is that reducing those costs will depend not only on cheaper models, but also on making the surrounding orchestration layer more efficient.
Featured image credits: Magnific.com
For more stories like it, click the +Follow button at the top of this page to follow us.