2025年4月最佳AI大模型排行

这是一份AI大模型评测报告，提供多个行业流行的测试基准评分，测试基准详见：Chatbot Arena 、MMLU(5-shot)、 Arena-Hard-Auto

排行	模型名	ArenaElo	CodingElo	Vision	ArenaHard	MMLU	投票	组织	许可	模型详情


	Gemini-2.5-Pro-Exp-03-25	1443	1422	1342	0	0	3474	Google	Proprietary
	ChatGPT-4o-latest (2025-03-26)	1408	1431	1312	0	0	2676	OpenAI	Proprietary
	Grok-3-Preview-02-24	1406	1411	0	0	92.7	8926	xAI	Proprietary
4	chocolate (Early Grok-3)	1402	1399	0	0	0	7829	xAI	Proprietary
5	GPT-4.5-Preview	1400	1408	0	0	0	8297	OpenAI	Proprietary
6	Gemini-2.0-Flash-Thinking-Exp-01-21	1385	1368	1280	0	0	13336	Google	Proprietary
7	Gemini-2.0-Pro-Exp-02-05	1379	1372	1252	0	0	11197	Google	Proprietary
8	ChatGPT-4o-latest (2025-01-29)	1377	1360	1276	0	0	10529	OpenAI	Proprietary
9	DeepSeek-V3-0324	1370	1387	0	0	0	2840	DeepSeek	MIT
10	DeepSeek-R1	1361	1362	0	0	0	5079	DeepSeek	MIT
11	Gemini-2.0-Flash-001	1356	1353	1243	0	0	9092	Google	Proprietary
12	Gemini-2.0-Flash-Exp	1355	1353	1257	0	0	22526	Google	Proprietary
13	o1-2024-12-17	1353	1363	0	90.4	0	15437	OpenAI	Proprietary
14	Gemma-3-27B-it	1338	1306	0	0	0	4794	Google	Gemma
15	Qwen2.5-Max	1332	1335	0	0	0	7370	Alibaba	Proprietary
16	o3-mini-high	1326	1364	0	0	0	12604	OpenAI	Proprietary
17	DeepSeek-V3	1317	1317	0	0	0	17717	DeepSeek	DeepSeek
18	Qwen2.5-Plus-0125	1313	1316	0	0	0	3682	Alibaba	Proprietary
19	QwQ-32B	1312	1329	0	0	0	2691	Alibaba	Apache 2.0
20	Command A (03-2025)	1312	1318	0	0	0	3000	Cohere	CC-BY-NC-4.0
21	Gemini-2.0-Flash-Lite-Preview-02-05	1310	1322	1153	0	0	8465	Google	Proprietary
22	Qwen-Plus-0125	1310	1320	0	0	0	6058	Alibaba	Proprietary
23	Gemini-2.0-Flash-Lite	1309	1319	1152	0	0	16354	Google	Proprietary
24	GLM-4-Plus-0111	1308	1291	0	0	0	4171	Zhipu	Proprietary
25	o3-mini	1305	1355	0	0	0	9338	OpenAI	Proprietary
26	o1-mini	1304	1353	0	92	0	55053	OpenAI	Proprietary
27	Step-2-16K-Exp	1304	1296	0	0	0	5133	StepFun	Proprietary
28	Hunyuan-TurboS-20250226	1303	1316	0	0	0	1443	Tencent	Proprietary
29	Gemini-1.5-Pro-002	1302	1290	1221	0	0	53271	Google	Proprietary
30	Claude 3.7 Sonnet (thinking-32k)	1302	1332	0	0	0	4487	Anthropic	Proprietary
31	Claude 3.7 Sonnet	1300	1342	1225	0	0	8411	Anthropic	Proprietary
32	Hunyuan-Turbo-0110	1296	1316	0	0	0	1473	Tencent	Proprietary
33	Llama-3.3-Nemotron-Super-49B-v1	1294	1293	0	88.3	0	2232	Nvidia	Nvidia
34	Grok-2-08-13	1288	1282	0	0	0	67117	xAI	Proprietary
35	Yi-Lightning	1287	1303	0	81.5	0	28955	01 AI	Proprietary
36	GPT-4o-2024-05-13	1285	1293	1206	79.21	88.7	117763	OpenAI	Proprietary
37	Claude 3.5 Sonnet (20241022)	1284	1325	1184	85.2	88.7	54384	Anthropic	Proprietary
38	Deepseek-v2.5-1210	1279	1297	0	0	0	7247	DeepSeek	DeepSeek
39	Athene-v2-Chat-72B	1275	1300	0	85	0	26182	NexusFlow	NexusFlow
40	GPT-4o-mini-2024-07-18	1273	1284	1124	74.94	82	64411	OpenAI	Proprietary
41	Hunyuan-Large-2025-02-10	1272	1293	0	0	0	3862	Tencent	Proprietary
42	Gemini-1.5-Flash-002	1271	1254	1205	0	0	37039	Google	Proprietary
43	Llama-3.1-405B-Instruct	1269	1280	0	0	88.6	29561	Meta	Llama 3.1
44	Llama-3.1-Nemotron-70B-Instruct	1268	1271	0	84.9	0	7598	Nvidia	Llama 3.1
45	Llama-3.1-405B-Instruct-bf16	1268	1280	0	0	88.6	37887	Meta	Llama 3.1
46	Llama-3.1-405B-Instruct-fp8	1267	1276	0	69.3	88.6	63074	Meta	Llama 3.1
47	Grok-2-Mini-08-13	1266	1262	0	0	0	55479	xAI	Proprietary
48	Yi-Lightning-lite	1264	1267	0	0	0	17062	01 AI	Proprietary
49	Hunyuan-Standard-2025-02-10	1260	1269	0	0	0	4022	Tencent	Proprietary
50	Qwen2.5-72B-Instruct	1257	1283	0	78	0	41575	Alibaba	Qwen
51	GPT-4-Turbo-2024-04-09	1256	1263	1151	82.63	0	102116	OpenAI	Proprietary
52	Llama-3.3-70B-Instruct	1255	1258	0	0	0	23945	Meta	Llama-3.3
53	Mistral-Large-2407	1252	1269	0	70.42	0	48191	Mistral	Mistral Research
54	GPT-4-1106-preview	1250	1253	0	0	0	103732	OpenAI	Proprietary
55	Athene-70B	1250	1253	0	77.6	0	20602	NexusFlow	CC-BY-NC-4.0
56	Llama-3.1-70B-Instruct	1248	1251	0	55.73	86	58756	Meta	Llama 3.1
57	Claude 3 Opus	1247	1250	1076	60.36	86.8	202670	Anthropic	Proprietary
58	Mistral-Large-2411	1246	1265	0	0	0	18879	Mistral	MRL
59	Amazon Nova Pro 1.0	1246	1260	1044	0	0	17108	Amazon	Proprietary
60	GPT-4-0125-preview	1245	1244	0	77.96	0	97040	OpenAI	Proprietary
61	Llama-3.1-Tulu-3-70B	1244	1231	0	0	0	3016	Ai2	Llama 3.1
62	Yi-Large-preview	1240	1245	0	71.48	0	51649	01 AI	Proprietary
63	Claude 3.5 Haiku (20241022)	1236	1263	0	0	0	19050	Anthropic	Propretary
64	Reka-Core-20240904	1235	1222	0	0	0	7933	Reka AI	Proprietary
65	Reka-Core-20240722	1231	1208	0	0	0	13288	Reka AI	Proprietary
66	Qwen-Plus-0828	1227	1245	0	0	0	14612	Alibaba	Proprietary
67	Gemini-1.5-Flash-001	1227	1232	1072	49.61	78.9	65662	Google	Proprietary
68	Jamba-1.5-Large	1221	1228	0	0	81.2	9123	AI21 Labs	Jamba Open
69	Deepseek-v2-API-0628	1220	1242	0	0	0	19500	DeepSeek AI	DeepSeek
70	Gemma-2-27B-it	1220	1210	0	57.51	0	76608	Google	Gemma license
71	Qwen2.5-Coder-32B-Instruct	1217	1261	0	0	0	5721	Alibaba	Apache 2.0
72	Amazon Nova Lite 1.0	1217	1234	1060	0	0	15012	Amazon	Proprietary
73	Mistral-Small-24B-Instruct-2501	1217	1233	0	0	0	3481	Mistral	Apache 2.0
74	Gemma-2-9B-it-SimPO	1216	1196	0	0	0	10553	Princeton	MIT
75	Command R+ (08-2024)	1215	1181	0	0	0	10538	Cohere	CC-BY-NC-4.0
76	Deepseek-Coder-v2-0724	1214	1266	0	62.3	0	11724	DeepSeek	Proprietary
77	Yi-Large	1213	1220	0	63.7	0	16632	01 AI	Proprietary
78	Gemini-1.5-Flash-8B-001	1213	1208	1106	0	0	37728	Google	Proprietary
79	Llama-3.1-Nemotron-51B-Instruct	1211	1211	0	0	0	3895	Nvidia	Llama 3.1
80	Nemotron-4-340B-Instruct	1209	1198	0	0	0	20605	Nvidia	NVIDIA Open Model
81	Aya-Expanse-32B	1209	1192	0	0	0	28806	Cohere	CC-BY-NC-4.0
82	Gemini App (2024-01-24)	1208	1171	0	0	0	11822	Google	Proprietary
83	GLM-4-0520	1207	1216	0	63.84	0	10215	Zhipu AI	Proprietary
84	Llama-3-70B-Instruct	1207	1200	0	46.57	82	163746	Meta	Llama 3
85	Phi-4	1205	1234	0	0	0	11133	Microsoft	MIT
86	Reka-Flash-20240904	1205	1191	0	0	0	8125	Reka AI	Proprietary
87	Gemini-1.5-Flash-8B-Exp-0827	1205	1189	1112	0	0	25356	Google	Proprietary
88	Claude 3 Sonnet	1201	1213	1048	46.8	79	113001	Anthropic	Proprietary
89	Reka-Flash-20240722	1201	1187	0	0	0	13727	Reka AI	Proprietary
90	Reka-Core-20240501	1200	1190	1015	0	83.2	62561	Reka AI	Proprietary
91	Amazon Nova Micro 1.0	1198	1210	0	0	0	15029	Amazon	Proprietary
92	Gemma-2-9B-it	1192	1173	0	0	0	54333	Google	Gemma license
93	Command R+ (04-2024)	1190	1164	0	33.07	0	80864	Cohere	CC-BY-NC-4.0
94	Hunyuan-Standard-256K	1189	1227	0	0	0	2898	Tencent	Proprietary
95	Qwen2-72B-Instruct	1187	1187	0	46.86	84.2	38877	Alibaba	Qianwen LICENSE
96	GPT-4-0314	1186	1196	0	50	86.4	55947	OpenAI	Proprietary
97	Llama-3.1-Tulu-3-8B	1185	1178	0	0	0	3071	Ai2	Llama 3.1
98	GLM-4-0116	1183	1191	0	55.72	0	7575	Zhipu AI	Proprietary
99	Qwen-Max-0428	1183	1190	0	0	0	25678	Alibaba	Proprietary
100	Ministral-8B-2410	1182	1201	0	0	0	5108	Mistral	MRL
101	Aya-Expanse-8B	1180	1164	0	0	0	10469	Cohere	CC-BY-NC-4.0
102	Command R (08-2024)	1180	1162	0	0	0	10843	Cohere	CC-BY-NC-4.0
103	Claude 3 Haiku	1179	1189	1000	41.47	75.2	122288	Anthropic	Proprietary
104	DeepSeek-Coder-V2-Instruct	1178	1239	0	0	0	15757	DeepSeek AI	DeepSeek License
105	Llama-3.1-8B-Instruct	1176	1186	0	21.34	73	52635	Meta	Llama 3.1
106	Jamba-1.5-Mini	1176	1181	0	0	69.7	9271	AI21 Labs	Jamba Open
107	Reka-Flash-Preview-20240611	1165	1155	1024	0	0	20410	Reka AI	Proprietary
108	GPT-4-0613	1163	1167	0	37.9	0	91617	OpenAI	Proprietary
109	Qwen1.5-110B-Chat	1161	1175	0	0	80.4	27448	Alibaba	Qianwen LICENSE
110	Mistral-Large-2402	1157	1170	0	37.71	81.2	64904	Mistral	Proprietary
111	Yi-1.5-34B-Chat	1157	1162	0	0	76.8	25130	01 AI	Apache-2.0
112	Reka-Flash-21B-online	1156	1147	0	0	0	16025	Reka AI	Proprietary
113	QwQ-32B-Preview	1153	1146	0	0	0	3413	Alibaba	Apache 2.0
114	Llama-3-8B-Instruct	1152	1146	0	20.56	68.4	109223	Meta	Llama 3
115	InternLM2.5-20B-chat	1149	1158	0	0	0	10591	InternLM	Other
116	Claude-1	1149	1136	0	0	77	21159	Anthropic	Proprietary
117	Command R (04-2024)	1149	1123	0	17.02	0	56361	Cohere	CC-BY-NC-4.0
118	Mixtral-8x22b-Instruct-v0.1	1148	1153	0	36.36	77.8	53777	Mistral	Apache 2.0
119	Mistral Medium	1148	1152	0	31.9	75.3	35554	Mistral	Proprietary
120	Reka-Flash-21B	1148	1141	0	0	73.5	25808	Reka AI	Proprietary
121	Qwen1.5-72B-Chat	1147	1160	0	36.12	77.5	40652	Alibaba	Qianwen LICENSE
122	Granite-3.1-8B-Instruct	1143	1172	0	0	0	3303	IBM	Apache 2.0
123	Gemma-2-2b-it	1143	1108	0	0	51.3	46048	Google	Gemma license
124	Claude-2.0	1132	1135	0	23.99	78.5	12761	Anthropic	Proprietary
125	Gemini-1.0-Pro-001	1131	1103	0	0	71.8	18788	Google	Proprietary
126	Zephyr-ORPO-141b-A35b-v0.1	1127	1124	0	0	0	4861	HuggingFace	Apache 2.0
127	Qwen1.5-32B-Chat	1125	1149	0	0	73.4	22762	Alibaba	Qianwen LICENSE
128	Mistral-Next	1124	1132	0	27.37	0	12374	Mistral	Proprietary
129	Phi-3-Medium-4k-Instruct	1123	1125	0	33.37	78	26105	Microsoft	MIT
130	Granite-3.1-2B-Instruct	1119	1146	0	0	0	3386	IBM	Apache 2.0
131	Starling-LM-7B-beta	1119	1129	0	23.01	0	16669	Nexusflow	Apache-2.0
132	Claude-2.1	1118	1132	0	22.77	0	37693	Anthropic	Proprietary
133	GPT-3.5-Turbo-0613	1117	1135	0	24.82	0	38944	OpenAI	Proprietary
134	Mixtral-8x7B-Instruct-v0.1	1114	1114	0	23.4	70.6	76125	Mistral	Apache 2.0
135	Claude-Instant-1	1111	1109	0	0	73.4	20619	Anthropic	Proprietary
136	Yi-34B-Chat	1111	1106	0	23.15	73.5	15921	01 AI	Yi License
137	Gemini Pro	1111	1092	0	17.8	71.8	6559	Google	Proprietary
138	Qwen1.5-14B-Chat	1109	1126	0	0	67.6	18670	Alibaba	Qianwen LICENSE
139	GPT-3.5-Turbo-0125	1106	1124	0	23.34	0	68857	OpenAI	Proprietary
140	GPT-3.5-Turbo-0314	1106	1115	0	18.05	70	5645	OpenAI	Proprietary
141	WizardLM-70B-v1.0	1106	1071	0	0	63.7	8383	Microsoft	Llama 2
142	DBRX-Instruct-Preview	1103	1118	0	24.63	73.7	33730	Databricks	DBRX LICENSE
143	Llama-3.2-3B-Instruct	1103	1080	0	0	0	8404	Meta	Llama 3.2
144	Phi-3-Small-8k-Instruct	1102	1107	0	29.77	75.7	18477	Microsoft	MIT
145	Tulu-2-DPO-70B	1099	1093	0	14.99	0	6661	AllenAI/UW	AI2 ImpACT Low-risk
146	Granite-3.0-8B-Instruct	1093	1097	0	0	0	7005	IBM	Apache 2.0
147	Llama-2-70B-chat	1093	1072	0	11.55	63	39616	Meta	Llama 2
148	OpenChat-3.5-0106	1092	1102	0	0	65.8	12977	OpenChat	Apache-2.0
149	Vicuna-33B	1091	1067	0	8.63	59.2	22950	LMSYS	Non-commercial
150	Snowflake Arctic Instruct	1090	1077	0	17.61	67.3	34168	Snowflake	Apache 2.0
151	Starling-LM-7B-alpha	1088	1080	0	12.8	63.9	10417	UC Berkeley	CC-BY-NC-4.0
152	Gemma-1.1-7B-it	1084	1084	0	0	64.3	25067	Google	Gemma license
153	Nous-Hermes-2-Mixtral-8x7B-DPO	1084	1079	0	0	0	3833	NousResearch	Apache-2.0
154	NV-Llama2-70B-SteerLM-Chat	1081	1023	0	0	68.5	3638	Nvidia	Llama 2
155	pplx-70B-online	1078	1028	0	0	0	6894	Perplexity AI	Proprietary
156	DeepSeek-LLM-67B-Chat	1077	1079	0	0	71.3	4986	DeepSeek AI	DeepSeek License
157	OpenChat-3.5	1076	1054	0	0	64.3	8106	OpenChat	Apache-2.0
158	Granite-3.0-2B-Instruct	1074	1088	0	0	0	7186	IBM	Apache 2.0
159	OpenHermes-2.5-Mistral-7B	1074	1058	0	0	0	5086	NousResearch	Apache-2.0
160	Mistral-7B-Instruct-v0.2	1072	1074	0	0	0	20048	Mistral	Apache-2.0
161	Phi-3-Mini-4K-Instruct-June-24	1071	1082	0	0	70.9	12811	Microsoft	MIT
162	Qwen1.5-7B-Chat	1070	1089	0	0	61	4866	Alibaba	Qianwen LICENSE
163	GPT-3.5-Turbo-1106	1068	1095	0	0	0	17028	OpenAI	Proprietary
164	Phi-3-Mini-4k-Instruct	1066	1086	0	0	68.8	21089	Microsoft	MIT
165	Llama-2-13b-chat	1063	1051	0	0	53.6	19732	Meta	Llama 2
166	SOLAR-10.7B-Instruct-v1.0	1062	1047	0	0	66.2	4290	Upstage AI	CC-BY-NC-4.0
167	Dolphin-2.2.1-Mistral-7B	1062	1024	0	0	0	1713	Cognitive Computations	Apache-2.0
168	WizardLM-13b-v1.2	1059	1026	0	0	52.7	7175	Microsoft	Llama 2
169	Llama-3.2-1B-Instruct	1054	1047	0	0	0	8528	Meta	Llama 3.2
170	Step-1o-Vision-32k (highres)	0	0	1180	0	0	0	StepFun	Proprietary
171	Qwen2.5-VL-72B-Instruct	0	0	1164	0	0	0	Alibaba	Qwen
172	Pixtral-Large-2411	0	0	1153	0	0	0	Mistral	MRL
173	Qwen-VL-Max-1119	0	0	1127	0	0	0	Alibaba	Proprietary
174	Step-1V-32K	0	0	1111	0	0	0	StepFun	Proprietary
175	Qwen2-VL-72b-Instruct	0	0	1110	0	0	0	Alibaba	Qwen
176	Molmo-72B-0924	0	0	1075	0	0	0	AI2	Apache 2.0
177	Pixtral-12B-2409	0	0	1072	0	0	0	Mistral	Apache 2.0
178	Llama-3.2-90B-Vision-Instruct	0	0	1069	0	0	0	Meta	Llama 3.2
179	InternVL2-26B	0	0	1067	0	0	0	OpenGVLab	MIT
180	Qwen2-VL-7B-Instruct	0	0	1053	0	0	0	Aliaba	Apache 2.0
181	Yi-Vision	0	0	1045	0	0	0	01 AI	Proprietary
182	Llama-3.2-11B-Vision-Instruct	0	0	1032	0	0	0	Meta	Llama 3.2
183	Hunyuan-Standard-Vision-2024-12-31	0	0	1064	0	0	0	Tencent	Proprietary
184	Aya-Vision-32B	0	0	1063	0	0	0	Cohere	CC-BY-NC-4.0

2025年4月最佳AI大模型排行

猫目

友情链接

热门推荐

关于

关注我们