Across the AI landscape, users are growing increasingly aware of just how pricey their deployments can get—and feeling a fresh push to trim expenses. While open-source models come with notably lower per-token costs, pinpointing the right model for a specific task remains a challenge. On Thursday, Writer, a company providing AI tools and agents for marketers, unveiled its new flagship model, Palmyra X6, designed to tackle that issue for its customers. Built as a post-training variant of Z.ai’s open-source model GLM-5.2, Writer claims the new system delivers deployment-ready capabilities at a much more affordable price. The company projects that this model, paired with upgrades to its harness infrastructure, could reduce customer costs by up to 50% for routine tasks. Alongside the new model, Writer also rolled out major enhancements to its standard agentic harness. Both features became available to Writer clients starting Thursday. “They want flattening costs, and it seems like nobody can deliver that.”

The new strategy zeroes in on complex, multi-step tasks, executing them faster and with fewer tokens. Writer views harness optimization as a key lever in making this happen. A recent paper from Writer researchers supports this approach, testing small tweaks in harness efficiency across a range of models. The findings showed that, in many instances, adjusting the harness proved a more dependable cost-reduction method than switching models, with expenses dropping by an average of 40% throughout their tests. “The harness is the one component whose efficiency multiplies across every model an organization runs—present and future,” the researchers noted. For Writer’s clients, the experience remains model-agnostic: Palmyra X6 will coexist with other Writer models or external models brought in via Azure or Amazon Bedrock. Yet Habib also sees the cost-cutting push fueling broader skepticism toward major AI labs, which have a financial incentive to boost token consumption.