Key facts
- Writer launched its new AI model, Palmyra X6, built on Z.ai's open-source GLM-5.2.
- The company also upgraded its agentic harness infrastructure.
- Writer estimates these changes will reduce customer costs by up to 50% for basic tasks.
- Palmyra X6 is designed to handle complex, multi-step tasks more efficiently with fewer tokens.
- Research by Writer found harness optimization can reduce costs by an average of 40%.
Writer has launched its new flagship AI model, Palmyra X6, and simultaneously upgraded its agentic harness infrastructure, aiming to address the escalating costs associated with AI deployments for enterprise clients. The company estimates that these combined advancements will lead to cost reductions of up to 50% for basic tasks.
Palmyra X6 is a post-training variation of Z.ai's open-source GLM-5.2 model, designed to offer deployment-ready capabilities at a lower price point. The upgrades to Writer's harness are intended to optimize the execution of complex, multi-step tasks, requiring fewer tokens and thus reducing overall expenses. Writer's CEO, May Habib, stated that enterprises are seeking stable costs rather than just benchmark chasing, and expressed a growing distrust of major AI labs that may incentivize higher token usage.
Recent research from Writer supports the significance of harness optimization, indicating that improvements in the harness can be a more reliable method for cost reduction than solely relying on model choice. The company's findings suggest that harness efficiency gains can multiply across all models an organization uses. Writer clients will have access to Palmyra X6 and the harness upgrades starting Thursday, with the platform remaining model-agnostic, allowing integration of other Writer models or external models via Azure or Amazon Bedrock.