Leaderboard
Speaker SIM
speaker-simRecord
- Category
- speech-tts
- Scored models
- 27
Scores
| Model | Score |
|---|---|
| — | 96.9 |
| maya-research/Veena | 92 |
| maya-research/Veena | 92 |
| k2-fsa/OmniVoice | 89 |
| nineninesix/qwen3_5-full-attn-only-14 | 86.7 |
| k2-fsa/OmniVoice | 83 |
| k2-fsa/OmniVoice | 80.1 |
| k2-fsa/OmniVoice | 80.1 |
| zai-org/GLM-TTS | 79.6 |
| — | 79.49 |
| — | 78.4 |
| FunAudioLLM/Fun-CosyVoice3-0.5B-2512 | 78 |
| FunAudioLLM/CosyVoice2-0.5B | 75.7 |
| — | 75.5 |
| FunAudioLLM/CosyVoice-300M | 75.5 |
| FunAudioLLM/CosyVoice-300M | 75.5 |
| — | 73.5 |
| OpenMOSS-Team/MOSS-TTS-Local-Transformer | 73.5 |
| — | 73.5 |
| — | 73.5 |
| — | 73.4 |
| — | 71.46 |
| — | 71.46 |
| Qwen/Qwen2.5-0.5B | 64.3 |
Remaining fields
Identity, launch, related records, and measured speed are shown above. This is the rest of the normalized record.
- category
- speech-tts
- id
- speaker-sim
- name
- Speaker SIM
- schema version
- local-ai-registry/v1
rows
| variant | conf | context | model id | org | rank | root | score |
|---|---|---|---|---|---|---|---|
| A.X-K2-Raon-Speech-21B-A3B | medium | Speaker Similarity | Unknown | KRAFTON | 1 | — | 96.9 |
| Veena | medium | Speaker Similarity | veena | maya-research | 2 | maya-research/Veena | 92 |
| Veena-GGUF | medium | Speaker Similarity | veena | Mungert | 3 | maya-research/Veena | 92 |
| g-omnivoice | medium | SIM | omnivoice | g-group-ai-lab | 4 | k2-fsa/OmniVoice | 89 |
| gepard-1.0 | low | SIM ↑ | qwen3-5-full-attn-only-14 | nineninesix | 5 | nineninesix/qwen3_5-full-attn-only-14 | 86.7 |
| VoiceTut-TTS | medium | Speaker similarity (cloning, cosine) | omnivoice | mohammedaly22 | 6 | k2-fsa/OmniVoice | 83 |
| omnivoice-hakka-community-1 | low | SIM-o | omnivoice | formospeech | 7 | k2-fsa/OmniVoice | 80.1 |
| omnivoice-hakka-community-1 | medium | SIM-o | omnivoice | formospeech | 8 | k2-fsa/OmniVoice | 80.1 |
| GLM-TTS | low | Open-source | Unknown | zai-org | 9 | zai-org/GLM-TTS | 79.6 |
| MOSS-TTSD-v1.0 | high | ZH - SIM | Unknown | OpenMOSS-Team | 10 | — | 79.49 |
| hindi-tts-190M | low | Speaker SIM ↑ | Unknown | 5ivatej | 11 | — | 78.4 |
| Fun-CosyVoice3-0.5B-2512 | high | Open-Source | Unknown | FunAudioLLM | 12 | FunAudioLLM/Fun-CosyVoice3-0.5B-2512 | 78 |
| CosyVoice2-0.5B | high | Open-Source | Unknown | FunAudioLLM | 13 | FunAudioLLM/CosyVoice2-0.5B | 75.7 |
| CosyVoice-300M-SFT | low | Open-Source | Unknown | FunAudioLLM | 14 | — | 75.5 |
| CosyVoice-300M | low | Open-Source | Unknown | FunAudioLLM | 15 | FunAudioLLM/CosyVoice-300M | 75.5 |
| CosyVoice-300M-Instruct | low | Open-Source | Unknown | FunAudioLLM | 16 | FunAudioLLM/CosyVoice-300M | 75.5 |
| MOSS-TTS | low | Open‑source | Unknown | OpenMOSS-Team | 17 | — | 73.5 |
| MOSS-TTS-Local-Transformer | low | Open‑source | Unknown | OpenMOSS-Team | 18 | OpenMOSS-Team/MOSS-TTS-Local-Transformer | 73.5 |
| MOSS-TTS-GGUF | low | Open‑source | Unknown | OpenMOSS-Team | 19 | — | 73.5 |
| MOSS-TTS-GGUF | low | Open‑source | Unknown | niobures | 20 | — | 73.5 |
| Raon-OpenTTS-1B | low | SIM ↑ | Unknown | KRAFTON | 21 | — | 73.4 |
| MOSS-TTS-GGUF | high | Seed-TTS-eval — Baseline (Hugging Face) EN SIM (%) | Unknown | OpenMOSS-Team | 22 | — | 71.46 |
| MOSS-TTS-GGUF | high | Seed-TTS-eval — Baseline (Hugging Face) EN SIM (%) | Unknown | niobures | 23 | — | 71.46 |
| VibeVoice-Realtime-0.5B | low | Speaker Similarity ↑ | qwen-qwen2-5-0-5b | microsoft | 24 | Qwen/Qwen2.5-0.5B | 64.3 |
| VibeVoice-Realtime-0.5B-ONNX | low | Speaker Similarity ↑ | qwen-qwen2-5-0-5b | elbruno | 25 | microsoft/VibeVoice-Realtime-0.5B | 64.3 |
| JaiTTS-F5TTS | high | Short SIM | Unknown | JTS-AI | 26 | SWivid/F5-TTS | 60 |
| f5-tts-quebec-french-finetune | medium | Speaker similarity ↑ (ECAPA cosine) | Unknown | tontate | 27 | RASPIAUDIO/F5-French-MixedSpeakers-reduced | 47.9 |
Provenance & metadata (1)
source
- kind
- leaderboard-scrape
- paths
- benchmarks/speaker-sim.html, docs/notes/hf-benchmark-contexts.json