Tool-calling on Confetti
72.7AST Soft AccuracyGemini-3.1 Flash-Live
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini-3.1 Flash-LiveTTS Model=Gemini-2.5-Pro, Voice=Orus (M)2026.05 | 72.7 | |
| Gemini-3.1 Flash-LiveTTS Model=GPT-4o-Mini, Voice=Coral (F)2026.05 | 71.2 | |
| Gemini-3.1 Flash-LiveTTS Model=Average2026.05 | 70.4 | |
| Gemini-3.1 Flash-LiveTTS Model=GPT-4o-Mini, Voice=Ash (M)2026.05 | 70.2 | |
| Gemini-3.1 Flash-LiveTTS Model=Gemini-2.5-Pro, Voice=Kore (F)2026.05 | 69.9 | |
| Gemini-3.1 Flash-LiveTTS Model=Gemini-2.5-Flash, Voice=Kore (F)2026.05 | 69.6 | |
| Gemini-3.1 Flash-LiveTTS Model=Gemini-2.5-Flash, Voice=Orus (M)2026.05 | 68.5 | |
| GPT-1.5 RealtimeTTS Model=GPT-4o-Mini, Voice=Ash (M)2026.05 | 64 | |
| GPT-1.5 RealtimeTTS Model=GPT-4o-Mini, Voice=Coral (F)2026.05 | 62.1 | |
| Qwen3-Omni 30B-A3BTTS Model=GPT-4o-Mini, Voice=Ash (M)2026.05 | 61.9 | |
| Qwen3-Omni 30B-A3BTTS Model=Gemini-2.5-Flash, Voice=Kore (F)2026.05 | 61.6 | |
| Qwen3-Omni 30B-A3BTTS Model=GPT-4o-Mini, Voice=Coral (F)2026.05 | 60.5 | |
| Qwen3-Omni 30B-A3BTTS Model=Average2026.05 | 60.4 | |
| Qwen3-Omni 30B-A3BTTS Model=Gemini-2.5-Pro, Voice=Kore (F)2026.05 | 60 | |
| GPT-1.5 RealtimeTTS Model=Gemini-2.5-Flash, Voice=Kore (F)2026.05 | 59.4 | |
| Qwen3-Omni 30B-A3BTTS Model=Gemini-2.5-Flash, Voice=Orus (M)2026.05 | 59.3 | |
| Qwen3-Omni 30B-A3BTTS Model=Gemini-2.5-Pro, Voice=Orus (M)2026.05 | 59.3 | |
| GPT-1.5 RealtimeTTS Model=Average2026.05 | 59.2 | |
| GPT-1.5 RealtimeTTS Model=Gemini-2.5-Pro, Voice=Kore (F)2026.05 | 58.8 | |
| GPT-1.5 RealtimeTTS Model=Gemini-2.5-Flash, Voice=Orus (M)2026.05 | 56.8 | |
| GPT-4o RealtimeTTS Model=GPT-4o-Mini, Voice=Coral (F)2026.05 | 55.9 | |
| GPT-4o RealtimeTTS Model=GPT-4o-Mini, Voice=Ash (M)2026.05 | 55.7 | |
| GPT-1.5 RealtimeTTS Model=Gemini-2.5-Pro, Voice=Orus (M)2026.05 | 54 | |
| GPT-4o RealtimeTTS Model=Average2026.05 | 51.2 | |
| GPT-4o RealtimeTTS Model=Gemini-2.5-Flash, Voice=Kore (F)2026.05 | 51 | |
| GPT-4o RealtimeTTS Model=Gemini-2.5-Flash, Voice=Orus (M)2026.05 | 50.3 | |
| GPT-4o RealtimeTTS Model=Gemini-2.5-Pro, Voice=Kore (F)2026.05 | 47.7 | |
| GPT-4o RealtimeTTS Model=Gemini-2.5-Pro, Voice=Orus (M)2026.05 | 46.7 | |
| GPT-Mini RealtimeTTS Model=GPT-4o-Mini, Voice=Coral (F)2026.05 | 45.5 | |
| GPT-Mini RealtimeTTS Model=GPT-4o-Mini, Voice=Ash (M)2026.05 | 42.9 | |
| GPT-Mini RealtimeTTS Model=Average2026.05 | 41.4 | |
| GPT-Mini RealtimeTTS Model=Gemini-2.5-Flash, Voice=Kore (F)2026.05 | 40.4 | |
| GPT-Mini RealtimeTTS Model=Gemini-2.5-Pro, Voice=Orus (M)2026.05 | 40.3 | |
| GPT-Mini RealtimeTTS Model=Gemini-2.5-Pro, Voice=Kore (F)2026.05 | 39.7 | |
| GPT-Mini RealtimeTTS Model=Gemini-2.5-Flash, Voice=Orus (M)2026.05 | 39.5 | |
| Gemini-2.5 Flash-LiveTTS Model=Gemini-2.5-Flash, Voice=Kore (F)2026.05 | 29.5 | |
| Gemini-2.5 Flash-LiveTTS Model=Gemini-2.5-Flash, Voice=Orus (M)2026.05 | 25.5 | |
| Phi-4 MMTTS Model=GPT-4o-Mini, Voice=Ash (M)2026.05 | 24.7 | |
| Phi-4 MMTTS Model=Gemini-2.5-Flash, Voice=Kore (F)2026.05 | 24.3 | |
| Phi-4 MMTTS Model=Gemini-2.5-Pro, Voice=Kore (F)2026.05 | 23.5 | |
| Phi-4 MMTTS Model=Gemini-2.5-Flash, Voice=Orus (M)2026.05 | 23.3 | |
| Phi-4 MMTTS Model=Average2026.05 | 23.3 | |
| Gemini-2.5 Flash-LiveTTS Model=Gemini-2.5-Pro, Voice=Orus (M)2026.05 | 23.1 | |
| Gemini-2.5 Flash-LiveTTS Model=Average2026.05 | 23.1 | |
| Phi-4 MMTTS Model=GPT-4o-Mini, Voice=Coral (F)2026.05 | 22.4 | |
| Gemini-2.5 Flash-LiveTTS Model=Gemini-2.5-Pro, Voice=Kore (F)2026.05 | 21.4 | |
| Phi-4 MMTTS Model=Gemini-2.5-Pro, Voice=Orus (M)2026.05 | 21.4 | |
| Gemini-2.5 Flash-LiveTTS Model=GPT-4o-Mini, Voice=Coral (F)2026.05 | 20 | |
| Gemini-2.5 Flash-LiveTTS Model=GPT-4o-Mini, Voice=Ash (M)2026.05 | 19.2 |