Rumble-2-Flash open-sourced: 0.5B speech model for 40+ languages
Reson8 open-sourced Rumble-2-Flash under Apache 2.0, a 0.5B-parameter speech recognition model supporting 40+ languages with automatic detection, competitive with models 3-60x its size. On a 302-hour call-center benchmark, it matches NVIDIA's Nemotron 3.5 Transcribe (30B) and processes one hour of audio in about half a second.