Writer has unveiled the Palmyra X6 AI model to reduce token costs
Writer, a company that develops AI-powered tools and agents for marketers, has unveiled its flagship model, the Palmyra X6, and updated its standard agent harness. Both new offerings became available to the company’s customers on Thursday, according to TechCrunch.
The Palmyra X6 was created as a variation of Z.ai’s open-source GLM-5.2 model following additional training. Writer stated that the model is intended to provide deployment-ready capabilities at a lower price. The company estimates that the combination of the Palmyra X6 and changes to its harness infrastructure could reduce customer costs for basic tasks by up to 50%.
The new approach focuses on complex, multi-step tasks that need to be performed faster and with lower token usage. Writer CEO May Habib told TechCrunch that enterprise customers are tired of chasing new benchmarks in model tests and are looking to stabilize their costs.
For more breaking news, follow the UA.News Telegram channel.
In a recent study, Writer researchers examined the impact of small changes in harness efficiency on various models. According to the test results, in many cases, optimizing the harness was a more reliable way to reduce costs than choosing a different model: the average cost reduction was 40%.
Writer’s customers will be able to use Palmyra X6 alongside the company’s other models or third-party models imported via Azure or Amazon Bedrock. Habib also stated that rising token costs are fueling IT leaders’ distrust of large AI labs.
