Across the AI industry, organizations are becoming increasingly aware of the high costs associated with their deployments—and feeling a renewed urgency to reduce them. While open-source models offer significantly lower per-token costs, finding the right model for a specific task remains a challenge.
On Thursday, Writer, a provider of AI tools and agents for marketers, launched its new flagship model, Palmyra X6, to address this problem. Built as a post-training variation of Z.ai's open-source model GLM-5.2, Writer claims the new system delivers deployment-ready capabilities at a fraction of the cost. The company estimates that Palmyra X6, combined with upgrades to its harness infrastructure, will reduce customer costs by up to 50% for basic tasks.
Alongside the new model, Writer released significant enhancements to its standard agentic harness. Both features are available to Writer clients starting Thursday. The company views harness optimization as a critical lever for cutting costs while maintaining performance on complex, multi-step tasks, which are executed faster and with fewer tokens.
“I think the enterprise is absolutely sick of chasing the next benchmark,” CEO May Habib told TechCrunch. “They want flattening cost, and it seems like nobody can deliver that.”
A recent paper by Writer researchers supports this approach, testing small adjustments in harness efficiency across multiple models. The research found that harness changes were often a more reliable cost-reduction method than model selection, achieving an average cost reduction of 40% across 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 integrated via Azure or Amazon Bedrock. Habib also sees the cost-cutting push as fueling a broader distrust toward major AI labs, which have a financial incentive to increase token usage.
“The cost explosion here is just unprecedented for customers, and so is the degree to which CIOs are giving up on the labs,” Habib told TechCrunch, adding that the AI labs “don't deeply understand how to help an enterprise get benefit from AI.”
via TechCrunch AI