Technology and Science13 August, 20262 min read
Writer Launches Palmyra X6 And Harness To Cut Enterprise AI Token Costs
This article is an automated aggregation of coverage from news outlets. Our editorials offer in-depth analysis curated by the NewsCord team.
Palmyra X6 flagship AI model released for enterprise marketing teams. New harness upgrades aim to cut token costs by about 50%.
Palmyra X6 cost push
Writer launched its flagship Palmyra X6 model alongside upgraded agent infrastructure called harness, positioning the release as a way to cut enterprise AI costs tied to token volume. Writer says Palmyra X6 is a post-trained variation of Z.ai’s open-weight GLM-5.2 and that, combined with harness changes, it can reduce the cost of completing basic tasks by up to 50%. In a comment to TechCrunch, CEO May Habib said, "I think the enterprise is absolutely sick of chasing the next benchmark," as she framed the push toward flattening costs.
TechCrunch also reported that Writer’s new approach emphasizes complex, multi-step tasks executed faster and with fewer tokens, with harness optimization presented as a crucial lever. Writer’s research, as quoted by TechCrunch, found that harness changes were often more reliable than model choice, with costs falling an average of 40% across testing.
Harness as the lever
Writer’s strategy centers on its agent harness, described as the infrastructure layer that manages requests, tool connections, and the sequence of actions performed by an AI agent while completing a task. In a TechCrunch account, Writer said the harness is the component whose efficiency multiplies across models, and the company’s researchers wrote that "The harness is the one component whose efficiency multiplies across every model an organization runs—present and future." The company also told TechCrunch that Palmyra X6 is model-agnostic for customers, with the model sitting alongside other Writer models or outside models imported through Azure or Amazon Bedrock.
May Habib linked rising token costs to distrust of major AI labs, telling TechCrunch that the labs "don’t deeply understand right how to help an enterprise get benefit from AI." Across the coverage, Writer’s harness optimization is presented as a way to deliver more consistent cost control without requiring enterprises to switch underlying models.
Pricing, governance, and adoption
Writer’s Palmyra X6 launch is tied to a broader enterprise push to manage token spending, with the company presenting the new model and harness upgrades as tools for stabilizing costs for customers. SiliconANGLE reported that Writer said its agent platform operates at an average of 52% lower cost, with a 48% improvement in speed and a 10% improvement in quality when paired with Palmyra X6. SiliconANGLE also quoted Waseem AlShikh saying, "The enterprise wants token consumption to explode — it means adoption is happening — but they need costs to flatten," tying the technical changes to budget pressure.
Beyond performance, SiliconANGLE said Writer launched upgrades to reporting and governance on the backend of its agentic controls, including a section for Playbooks and Skills that lets administrators see usage data, manage status and ownership, and monitor performance and costs. In the same SiliconANGLE coverage, the company described the ability for administrators to set consumption levels and alert limits to specific thresholds, framing governance as part of how enterprises control spend as agents run unattended.