模型排行

排名	变化	模型	得分	置信区间
1	—	Sonic 3.6闭源Cartesia	1,282	±17
2	—	Realtime TTS-2闭源Inworld	1,252	±18
3	—	Qwen-Audio-3.0-TTS-Plus闭源Alibaba	1,241	±14
4	—	Simba 3.2闭源SpeechifyAI	1,240	±14
5	—	Luna TTS闭源VUI Labs	1,228	±14
6	—	Realtime TTS-2 Flash - Research Preview闭源Inworld	1,222	±15
7	—	Breeze TTS 2开源BreezeBlue	1,215	±17
8	—	v3 Conversational闭源ElevenLabs	1,210	±15
9	—	Gemini 3.1 Flash TTS闭源Google	1,208	±12
10	—	StepAudio 2.5 TTS (Aug 2026)闭源StepFun	1,205	±17
11	—	Sonic 3.5闭源Cartesia	1,193	±12
12	—	Lightning V3.1 Pro (Jul 2026)闭源Smallest.ai	1,190	±14
13	—	Soniox TTS Real-Time v2闭源Soniox	1,179	±11
14	—	Eleven v3闭源ElevenLabs	1,175	±11
15	—	Speech 2.8 HD闭源MiniMax	1,171	±11
16	—	Falcon 2闭源Murf AI	1,158	±15
17	—	Speech 2.8 Turbo闭源MiniMax	1,152	±11
18	—	Gradium TTS (Aug 2026)闭源Gradium	1,149	±15
19	—	Async Flash v1.5闭源Async	1,141	±13
20	—	Fish Audio S2.1 Pro闭源Fish Audio	1,141	±13
21	—	Step TTS 2 (Mar 2026)闭源StepFun	1,138	±12
22	—	Speech 2.6 HD闭源MiniMax	1,134	±11
23	—	Lightning V3.1 Pro TTS (Jun 2026)闭源Smallest.ai	1,133	±13
24	—	Azure HD 2.5闭源Microsoft	1,131	±12
25	—	Fish Audio S2 Pro开源Fish Audio	1,128	±12
26	—	Speech 2.6 Turbo闭源MiniMax	1,127	±11
27	—	SpaceXAI TTS闭源SpaceXAI	1,123	±16
28	—	Simba 3.0闭源SpeechifyAI	1,122	±12
29	—	Async Pro v1.0闭源Async	1,121	±13
30	—	Speech-02-HD闭源MiniMax	1,112	±11
31	—	TTS-1 HD闭源OpenAI	1,106	±11
32	—	Step Audio EditX (Mar 2026)开源StepFun	1,104	±13
33	—	Turbo v2.5闭源ElevenLabs	1,100	±10
34	—	Multilingual v2闭源ElevenLabs	1,099	±10
35	—	Chatterbox HD闭源Resemble AI	1,093	±12
36	—	TTS-1闭源OpenAI	1,091	±10
37	—	Gemini 2.5 Flash Lite TTS闭源Google	1,084	±12
38	—	Speech 2.6闭源Hithink	1,084	±13
39	—	Speech-02-Turbo闭源MiniMax	1,082	±11
40	—	Studio闭源Google	1,081	±11
41	—	OpenAudio S1闭源Fish Audio	1,080	±11
42	—	Flash v2.5闭源ElevenLabs	1,080	±11
43	—	Voxtral TTS开源Mistral	1,078	±13
44	—	Journey闭源Google	1,075	±13
45	—	Maya 2 Flash闭源Maya Research	1,072	±14
46	—	Sonic 3闭源Cartesia	1,067	±11
47	—	GPT-Realtime-2闭源OpenAI	1,067	±13
48	—	T2A-01-HD闭源MiniMax	1,066	±11
49	—	Magpie-Multilingual 357M (Feb 2026)开源NVIDIA	1,065	±13
50	—	Polly Generative闭源Amazon	1,063	±11
51	—	SIMBA 1.6闭源SpeechifyAI	1,062	±12
52	—	Kokoro 82M v1.0开源Kokoro	1,060	±11
53	—	Maya 2 Global闭源Maya Research	1,059	±14
53	—	MiMo-V2.5-TTS闭源Xiaomi	1,059	±13
55	—	Coda闭源Rime	1,058	±12
56	—	Chirp 3: HD闭源Google	1,057	±11
57	—	Octave 2闭源Hume AI	1,056	±12
58	—	Gemini 2.5 Flash TTS (Dec 2025)闭源Google	1,056	±12
59	—	OpenAudio S1 Mini闭源Fish Audio	1,050	±19
60	—	Async Flash v1.0闭源Async	1,048	±11
61	—	Polly Long-Form闭源Amazon	1,044	±13
62	—	Maya1开源Maya Research	1,043	±12
63	—	Higgs Audio V3 TTS开源Boson AI	1,038	±13
64	—	SIMBA 1.0闭源SpeechifyAI	1,034	±11
65	—	MAI-Voice-1闭源Microsoft	1,034	±12
66	—	Bland Speech v3闭源Bland AI	1,034	±13
67	—	T2A-01-Turbo闭源MiniMax	1,033	±11
68	—	Gemini 2.5 Pro (Dec 2025)闭源Google	1,031	±12
69	—	Octave TTS闭源Hume AI	1,029	±11
70	—	Azure Neural闭源Microsoft	1,029	±19
71	—	MiMo-V2-TTS闭源Xiaomi	1,024	±13
72	—	Chatterbox开源Resemble AI	1,021	±11
73	—	Lightning v3.1闭源Smallest.ai	1,021	±12
74	—	Raon SpeechLM闭源Krafton	1,009	±13
75	—	Arcana v3闭源Rime	1,005	±13
76	—	Zonos-v0.1开源Zyphra	1,000	±0
77	—	Murf Speech Gen 2闭源Murf AI	981	±11
78	—	VibeVoice 7B开源Microsoft	969	±13
79	—	VibeVoice 1.5B闭源Microsoft	966	±13
80	—	OpenVoice v2开源OpenVoice	955	±13
81	—	Magpie Multilingual闭源NVIDIA	946	±14
82	—	Qwen3 TTS Flash闭源Alibaba	945	±14
83	—	Neuphonic TTS闭源Neuphonic	937	±13
84	—	Kugel 3闭源KugelAudio	926	±13
85	—	Qwen3 TTS闭源Alibaba	926	±13
86	—	XTTS v2开源Coqui	922	±14
87	—	WaveNet闭源Google	915	±12
88	—	Mist V2闭源Rime	901	±14
89	—	StyleTTS 2开源StyleTTS	893	±15
90	—	Neural2闭源Google	891	±12
91	—	Polly Neural闭源Amazon	889	±14
92	—	Standard闭源Google	882	±12
93	—	Noiz TTS闭源Noiz	869	±15
94	—	MetaVoice v1开源MetaVoice	840	±17
95	—	Polly Standard闭源Amazon	825	±15

其余 45 个模型

智力进步

每家厂商当时最强的模型在 Artificial Analysis 智力指数上的分数，按模型发布的月份画。

涨得最多 · Anthropic +57.2目前最高 · Anthropic 64.6最高与最低相差 18.1 分时间范围 2025年11月 – 2026年11月

64.6Anthropic60.0Google56.9OpenAI50.1Meta46.5Kimi