模型排行

排名	变化	模型	得分	置信区间
1	—	Eleven v4闭源ElevenLabs	1,321	±18
2	—	Qwen-Audio-3.1-TTS-Plus闭源Alibaba	1,292	±18
3	—	Sonic 3.6闭源Cartesia	1,278	±16
4	—	Gemini 3.8 Flash TTS闭源Google	1,275	±16
5	—	Qwen-Audio-3.0-TTS-Plus闭源Alibaba	1,261	±15
6	—	Realtime TTS-2闭源Inworld	1,251	±16
7	—	Simba 3.2闭源SpeechifyAI	1,242	±13
8	—	Gemini 3.8 Flash-Lite TTS闭源Google	1,242	±15
9	—	Luna TTS闭源VUI Labs	1,227	±13
10	—	Realtime TTS-2 Flash闭源Inworld	1,214	±14
11	—	StepAudio 2.5 TTS (Aug 2026)闭源StepFun	1,210	±15
12	—	Gemini 3.1 Flash TTS闭源Google	1,209	±12
13	—	v3 Conversational闭源ElevenLabs	1,203	±14
14	—	Sonic 3.5闭源Cartesia	1,190	±12
15	—	Soniox TTS Real-Time v2闭源Soniox	1,179	±11
16	—	Eleven v3闭源ElevenLabs	1,174	±11
17	—	Speech 2.8 HD闭源MiniMax	1,173	±11
18	—	Lightning V3.1 Pro (Jul 2026)闭源Smallest.ai	1,171	±14
19	—	Falcon 2闭源Murf AI	1,155	±14
20	—	Speech 2.8 Turbo闭源MiniMax	1,153	±11
21	—	Gradium TTS (Aug 2026)闭源Gradium	1,151	±14
22	—	Fish Audio S2.1 Pro闭源Fish Audio	1,141	±13
23	—	Async Flash v1.5闭源Async	1,136	±13
24	—	SpaceXAI TTS闭源SpaceXAI	1,136	±15
25	—	Step TTS 2 (Mar 2026)闭源StepFun	1,134	±12
26	—	Azure HD 2.5闭源Microsoft	1,131	±12
27	—	Speech 2.6 HD闭源MiniMax	1,128	±11
28	—	Simba 3.0闭源SpeechifyAI	1,125	±12
29	—	Speech 2.6 Turbo闭源MiniMax	1,124	±11
30	—	Async Pro v1.0闭源Async	1,119	±13
31	—	Speech-02-HD闭源MiniMax	1,114	±11
32	—	Speech 2.6闭源Hithink	1,103	±13
33	—	TTS-1 HD闭源OpenAI	1,103	±12
34	—	Chatterbox HD闭源Resemble AI	1,102	±13
35	—	Turbo v2.5闭源ElevenLabs	1,098	±11
36	—	Multilingual v2闭源ElevenLabs	1,097	±10
37	—	TTS-1闭源OpenAI	1,090	±11
38	—	Speech-02-Turbo闭源MiniMax	1,090	±12
39	—	Gemini 2.5 Flash Lite TTS闭源Google	1,083	±12
40	—	OpenAudio S1闭源Fish Audio	1,082	±11
41	—	Studio闭源Google	1,079	±11
42	—	Flash v2.5闭源ElevenLabs	1,078	±11
43	—	GPT-Realtime-2闭源OpenAI	1,077	±13
44	—	Journey闭源Google	1,074	±13
45	—	Maya 2 Flash闭源Maya Research	1,071	±14
46	—	T2A-01-HD闭源MiniMax	1,066	±11
47	—	Polly Generative闭源Amazon	1,065	±12
48	—	Coda闭源Rime	1,062	±13
49	—	Sonic 3闭源Cartesia	1,060	±12
50	—	Gemini 2.5 Flash TTS (Dec 2025)闭源Google	1,058	±12
51	—	Chirp 3: HD闭源Google	1,056	±12
52	—	MiMo-V2.5-TTS闭源Xiaomi	1,055	±13
53	—	Async Flash v1.0闭源Async	1,052	±11
54	—	Octave 2闭源Hume AI	1,051	±12
55	—	Maya 2 Global闭源Maya Research	1,049	±14
56	—	Bland Speech v3闭源Bland AI	1,037	±13
57	—	Octave TTS闭源Hume AI	1,036	±11
58	—	Azure Neural闭源Microsoft	1,035	±19
59	—	MAI-Voice-1闭源Microsoft	1,034	±13
60	—	Polly Long-Form闭源Amazon	1,033	±13
61	—	Gemini 2.5 Pro (Dec 2025)闭源Google	1,033	±13
62	—	T2A-01-Turbo闭源MiniMax	1,031	±11
63	—	Lightning v3.1闭源Smallest.ai	1,024	±13
63	—	MiMo-V2-TTS闭源Xiaomi	1,024	±14
65	—	Raon SpeechLM闭源Krafton	1,003	±14
66	—	Murf Speech Gen 2闭源Murf AI	984	±12
67	—	Qwen3 TTS Flash闭源Alibaba	944	±14
68	—	Neuphonic TTS闭源Neuphonic	940	±14
69	—	Kugel 3闭源KugelAudio	935	±13
70	—	Qwen3 TTS闭源Alibaba	932	±13
71	—	WaveNet闭源Google	918	±12
72	—	Mist V2闭源Rime	909	±15
73	—	Neural2闭源Google	895	±12
74	—	Polly Neural闭源Amazon	894	±14
75	—	Standard闭源Google	886	±12
76	—	Noiz TTS闭源Noiz	885	±15
77	—	Polly Standard闭源Amazon	830	±16

其余 27 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