ModelCards
← All benchmarks

960 ms delay

We haven't written a description for this benchmark yet; see the card notes below for how each maker ran it.

Your workloadRAG over docs · 12K in · 600 out · 50K requests/mo · 30% cached
  1. 1Voxtral Mini Transcribe Realtime 3 · Mistral AI
    6.0%

“Best value”: no other model here reports a higher score at a lower cost for your workload. Scores are self-reported by each maker with different settings, so treat close calls as ties.

How each maker reported it

  • Voxtral Mini Transcribe Realtime 3 launch post (FLEURS 35 chart)

    Mistral's own evaluation: macro-average WER (CER for ja, ko, zh) over 35 FLEURS languages. Voxtral Mini Transcribe 3 (offline) is shown as a baseline at 4.0%. Competitors in the same chart: Gemini 3.5 Realtime Transcribe 4.2%, Microsoft AI Transcribe 2 Streaming 5.8%, GPT Live Transcribe 8.4%, Scribe v2 Realtime 8.7%. Lower is better.

Only each model's own maker's number is shown; scores a maker reports for rival models are left out.