模型排行

排名	变化	模型	得分	置信区间
1	—	Eleven v4 Turbo闭源ElevenLabs	1,332	±17
2	—	Eleven v4闭源ElevenLabs	1,329	±16
3	—	Qwen-Audio-3.1-TTS-Plus闭源Alibaba	1,297	±16
4	—	Sonic 3.6闭源Cartesia	1,291	±14
5	—	Gemini 3.8 Flash TTS闭源Google	1,280	±14
6	—	Qwen-Audio-3.0-TTS-Plus闭源Alibaba	1,264	±14
7	—	Realtime TTS-2闭源Inworld	1,262	±15
8	—	Gemini 3.8 Flash-Lite TTS闭源Google	1,252	±14
9	—	Simba 3.2闭源SpeechifyAI	1,245	±12
10	—	StepAudio 2.5 TTS (Aug 2026)闭源StepFun	1,217	±14
11	—	Realtime TTS-2 Flash闭源Inworld	1,217	±13
12	—	Luna TTS闭源VUI Labs	1,213	±12
13	—	Gemini 3.1 Flash TTS闭源Google	1,212	±11
14	—	v3 Conversational闭源ElevenLabs	1,212	±13
15	—	Sonic 3.5闭源Cartesia	1,191	±11
16	—	Fish Audio S2.1 Pro (Oct 2026)闭源Fish Audio	1,187	±14
17	—	Eleven v3闭源ElevenLabs	1,182	±11
18	—	Soniox TTS Real-Time v2闭源Soniox	1,179	±11
19	—	Gradium TTS (Sep 2026)闭源Gradium · 另有 1 个档位	1,175	±13
20	—	Speech 2.8 HD闭源MiniMax	1,174	±11
21	—	Lightning V3.1 Pro (Jul 2026)闭源Smallest.ai	1,168	±12
22	—	Speech 2.8 Turbo闭源MiniMax	1,155	±11
23	—	Falcon 2闭源Murf AI	1,153	±13
24	—	SpaceXAI TTS闭源SpaceXAI	1,149	±13
25	—	Azure HD 2.5闭源Microsoft	1,140	±12
26	—	Async Flash v1.5闭源Async	1,139	±12
27	—	Step TTS 2 (Mar 2026)闭源StepFun	1,135	±12
28	—	Speech 2.6 HD闭源MiniMax	1,129	±11
29	—	Async Pro v1.0闭源Async	1,127	±12
30	—	Simba 3.0闭源SpeechifyAI	1,121	±11
31	—	Speech 2.6 Turbo闭源MiniMax	1,121	±11
32	—	Speech-02-HD闭源MiniMax	1,111	±11
33	—	Chatterbox HD闭源Resemble AI	1,109	±12
34	—	TTS-1 HD闭源OpenAI	1,102	±11
35	—	Turbo v2.5闭源ElevenLabs	1,098	±10
36	—	Speech 2.6闭源Hithink	1,096	±12
37	—	Multilingual v2闭源ElevenLabs	1,096	±10
38	—	Speech-02-Turbo闭源MiniMax	1,091	±11
39	—	TTS-1闭源OpenAI	1,089	±10
40	—	Gemini 2.5 Flash Lite TTS闭源Google	1,086	±11
41	—	OpenAudio S1闭源Fish Audio	1,085	±11
42	—	Studio闭源Google	1,077	±11
43	—	Flash v2.5闭源ElevenLabs	1,077	±11
44	—	Journey闭源Google	1,077	±12
45	—	GPT-Realtime-2闭源OpenAI	1,076	±12
46	—	T2A-01-HD闭源MiniMax	1,067	±11
47	—	Polly Generative闭源Amazon	1,066	±11
48	—	Maya 2 Flash闭源Maya Research	1,065	±13
49	—	Gemini 2.5 Flash TTS (Dec 2025)闭源Google	1,062	±12
50	—	Chirp 3: HD闭源Google	1,061	±12
51	—	Sonic 3闭源Cartesia	1,056	±12
52	—	Coda闭源Rime	1,056	±12
53	—	Octave 2闭源Hume AI	1,055	±12
54	—	MiMo-V2.5-TTS闭源Xiaomi	1,052	±12
55	—	Flux TTS闭源Deepgram	1,051	±13
56	—	Async Flash v1.0闭源Async	1,050	±11
57	—	Maya 2 Global闭源Maya Research	1,048	±13
58	—	Octave TTS闭源Hume AI	1,038	±11
59	—	Gemini 2.5 Pro (Dec 2025)闭源Google	1,038	±12
60	—	Bland Speech v3闭源Bland AI	1,037	±13
61	—	Azure Neural闭源Microsoft	1,034	±20
62	—	T2A-01-Turbo闭源MiniMax	1,030	±11
63	—	MAI-Voice-1闭源Microsoft	1,030	±12
64	—	Polly Long-Form闭源Amazon	1,029	±13
65	—	Lightning v3.1闭源Smallest.ai	1,022	±12
66	—	Raon SpeechLM闭源Krafton	1,005	±13
67	—	Murf Speech Gen 2闭源Murf AI	986	±12
68	—	Neuphonic TTS闭源Neuphonic	943	±14
69	—	Qwen3 TTS Flash闭源Alibaba	942	±14
70	—	Qwen3 TTS闭源Alibaba	930	±13
71	—	Kugel 3闭源KugelAudio	930	±13
72	—	WaveNet闭源Google	918	±12
73	—	Mist V2闭源Rime	904	±15
74	—	Polly Neural闭源Amazon	897	±14
75	—	Neural2闭源Google	894	±12
76	—	Standard闭源Google	886	±13
77	—	Noiz TTS闭源Noiz	875	±15
78	—	Polly Standard闭源Amazon	831	±16

其余 28 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