Writer launches Palmyra X6 model and upgraded harness to halve token costs
Writer introduced its new Palmyra X6 AI model and a revamped agentic harness, promising up to 50% lower token expenses for customers.
Writer unveiled Palmyra X6, a new flagship AI model built as a post-training adaptation of Z.ai’s open-source GLM-5.2, aimed at delivering lower per-token costs for enterprise users. The launch coincides with a substantial upgrade to Writer’s agentic harness, which the firm claims can slash token usage by up to half for basic operations. CEO May Habib emphasized that enterprises are tired of chasing benchmarks and need predictable, reduced expenses, noting that major AI labs often drive up token consumption.
Writer’s own research paper found that optimizing the harness can lower costs by an average of 40%, sometimes more than changing the underlying model. Customers can deploy Palmyra X6 alongside existing Writer models or external models hosted on Azure or Amazon Bedrock. The combined rollout is intended to make complex, multi-step AI tasks faster and cheaper for businesses.
Why it matters
Lower AI token costs could make advanced language models more affordable for enterprises.
In this story
