ModelCards
← All benchmarks

French (fr-FR)

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. 1Sonic 3.6 · Cartesia
    86.0%

“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

  • Sonic 3.6 launch post

    Cartesia's own claims. Elo values are Cartesia's reading of the Artificial Analysis TTS leaderboards (controlled voice and provider voices), not checked against the leaderboard here. Vote shares are from Cartesia's Aug 2026 native-speaker blind tests vs Eleven v3 (vote share = wins ÷ valid votes; ties make up the rest).

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