Fish Audio released Drama 3, a 35-billion-parameter mixture-of-experts text-to-speech model with approximately 3 billion active parameters. The model accepts plain-language direction, supports 83 languages, and ships under an open-weight community license with revenue caps (MarkTechPost). Separately, TypeSafe AI released Jev 1.14 under the jev-preview alias, adding experimental image input support, a Box bounding-box type, and optional 262,000-token context window at a rate of $0.042 per million input tokens (MarkTechPost).
In related model releases, SpaceXAI launched Grok Voice 3.0, a speech-to-speech model that ranks first on the Artificial Analysis Speech-to-Speech Index with a score of 86.9, priced at $0.08 per minute (MarkTechPost). Datalab released OmniParseBench, an open OCR benchmark containing 16,288 pass-fail tests across 2,937 pages in 92 languages, enabling language and script-level performance analysis (MarkTechPost).