The October 7, 2026 announcement says the same model weights also transcribe English, French, Spanish, and Portuguese, without a language flag, and provide word-level timestamps.
TII reports an average 20.92% word error rate across six Arabic test sets in the Open Universal Arabic ASR Leaderboard protocol. It says this was 2.25 percentage points below the best published result in its September 30 leaderboard snapshot. On TII's internal Emirati and Gulf-speech evaluation, Falcon-ASR recorded 22.73% WER and 10.19% character error rate; TII says its WER was 4.07 points below Qwen3-Omni in that comparison.
A Hugging Face demo is available, including transcription of clips up to 60 seconds. API access and native applications are planned, not announced as available. The source does not establish independent benchmark replication, performance outside its stated test sets, model licensing, or production availability.