Add results from Hugging Face's Open TTS Leaderboard

README.md

@@ -204,6 +204,24 @@ Voice variants sharing the same weights are merged. Inflect totals include the i
 
 </details>
 
+### 6. Open TTS Leaderboard
+
+As of **September 9, 2026**, on Hugging Face's [Open TTS Leaderboard](https://huggingface.co/spaces/hf-audio/open_tts_leaderboard) — which scores open-source TTS models on the same texts across two multilingual eval sets ([Seed-TTS eval](https://github.com/BytedanceSpeech/seed-tts-eval) and [CV3-Eval](https://github.com/QwenAudio/CV3-Eval)) — Inflect-Micro-v2 is:
+
+- 📈 On the **Pareto frontier** for **Model size vs WER**.
+- 🥉 **3rd on RTFx** (generation speed).
+
+<p align="center">
+  <img width="45%" alt="Model size vs WER Pareto frontier" src="https://huggingface.co/datasets/bezzam/tts_leaderboard_screenshots/resolve/main/inflect_micro-v2/wer_size.png" />
+  <img width="45%" alt="Top models by RTFx" src="https://huggingface.co/datasets/bezzam/tts_leaderboard_screenshots/resolve/main/inflect_micro-v2/top_rtfx.png" />
+</p>
+
+Screenshots from the "Streaming" tab, namely batch size 1 on an H200 GPU: 🥇 **1st on RTFx**, and 7th on TTFA (time-to-first-audio).
+<p align="center">
+  <img width="45%" alt="Top models by RTFx (streaming, GPU H200)" src="https://huggingface.co/datasets/bezzam/tts_leaderboard_screenshots/resolve/main/inflect_micro-v2/streaming_top_rtfx.png" />
+  <img width="45%" alt="Top models by TTFA (streaming, GPU H200)" src="https://huggingface.co/datasets/bezzam/tts_leaderboard_screenshots/resolve/main/inflect_micro-v2/streaming_top_ttfa.png" />
+</p>
+
 ---
 
 ## Choose the right Inflect