Synthio Labs released the DOSE (Drug-name Oral Synthesis Evaluation) benchmark on September 17, 2026, from San Francisco, calling it the first public benchmark built specifically to test how accurately voice AI models pronounce medication names. DOSE ran nine text-to-speech models against 274 drug names, including 146 newly approved ones, with each name read aloud inside a real clinical sentence rather than in isolation. Every model tested saw its pass rate drop on newly approved names, in some cases sharply, and leading voice AI models mispronounce up to one in three newly approved drug names. ElevenLabs' eleven_v3 saw a 26-point drop on new drug names, Google's Gemini TTS fell to 61.6%, and Microsoft Azure struggled most with generic drug names. The results were picked up by outlets including Yahoo Finance Singapore, extremetech.com, PYMNTS.com, Inshorts, finance.biggo.com, and NDTV Profit.