Pika unveils new generative media models and API Club

On August 18, 2026 Pika announced a series of new generative audio models—Pika Soundtrack, Music, SFX, and Speech—available through its API Club. The Soundtrack model generates synchronized, motion‑aware soundscapes from video; the Music model creates full tracks from text, lyrics, reference voice or existing audio; the SFX model produces focused sound effects from textual descriptions; and the Speech model is a 3‑billion‑parameter flow‑matching transformer that delivers studio‑quality 48 kHz speech, can clone a voice from seconds of reference audio, and runs at a 0.02 real‑time factor, according to Pika. In a separate post on September 17, 2026 the company introduced “the New Pika,” a unified platform that brings video, image, and audio creation together under a single interface, leveraging the company’s leading generative media models. Earlier, on August 5, 2026 Pika launched the Pika API Club, a membership service that provides access to more than 100 video, image, audio and LLM models, with pricing up to 88 % cheaper than competing API aggregators. Pika also claims its audio family is up to 20 × cheaper than comparable models. The announcements were made on the Pika blog; no independent testing or third‑party reviews were cited.