Artificial Analysis Openness Index

Name: Artificial Analysis Openness Index：得分
Creator: Artificial Analysis
License: https://artificialanalysis.ai/docs/legal/Terms-of-Use.pdf

A composite measure providing an industry standard to communicate model openness for users and developers.

查看示例任务

Artificial Analysis Openness Index 根据模型各组成部分的可用性和透明度，例如模型权重、训练数据和模型架构，评估模型的“开放”程度。

可用性表示能否通过 API 使用模型、能否利用开放权重自行托管，以及能否依据宽松许可自由使用。透明度衡量模型的方法和数据在多大程度上得到披露、共享并采用宽松许可，使社区能够了解模型输入、复现其方法或在此基础上继续开发。

所有评测均由 Artificial Analysis 独立进行。更多信息请参阅我们的 Artificial Analysis Openness Index 方法论。

Open Source Models

Openness Index

K2 Think V2在 Openness Index 中得分最高，为 89；其次是K2-V2 (High)，得分 89和K2-V2 (Medium)，得分 89

Openness Index

Gemini 3 Pro Preview (High)在 Openness Index 中得分最低，为 6；其次是GPT-5 Codex (High)，得分 6和GPT-5.1 (High)，得分 6

Openness Index

Artificial Analysis Openness Index：得分

Openness Index 以 0 至 100 的标准化评分衡量模型的开放程度（分数越高，越开放）

指数组成

Artificial Analysis Openness Index：组成部分

各组成部分对 Openness Index 的贡献，最高 18 分（分数越高，越开放）

透明度与可用性

Artificial Analysis Openness Index：模型可用性与透明度

模型可用性反映模型的使用途径及其许可（最高 6 分） · 模型透明度反映方法、数据披露与共享，以及训练过程相关的代码和许可（最高 12 分）

Most attractive quadrant

Openness Index 与发布日期

Artificial Analysis Openness Index：得分与发布日期

Artificial Analysis Openness Index · 发布日期

Most attractive region

开放性与智能水平

Artificial Analysis Openness Index 与 Artificial Analysis Intelligence Index

Most attractive quadrant

Openness Index 的构成

详细评估方法

维度

组成部分

评分

1. 模型可用性

权重

访问

0封闭权重，无 API

1封闭权重，API 限制 token 可见性

2封闭权重，API 可用

3开放权重

许可证

0封闭权重或不可商用

1可商用，需署名

2可商用，无需署名

3可商用，无需署名，无实质性限制

2. 模型透明度

数据：预训练与后训练（得分为各项的平均值）

访问

0未披露或有限披露

1部分数据来源详情及分类已披露

2完整数据配比已披露，大量数据已共享¹

3全部数据已共享

许可证（最严格的）

0不可商用/无大量数据共享

1可商用，需署名

2可商用，无需署名

3可商用，无需署名，无实质性限制

方法论

披露

0未披露或有限披露

1模型架构已披露

2有限的通用技术披露

3完整技术细节已披露

许可证（最严格的）

0未披露或发布任何代码

1框架已披露，可公开商用

2端到端训练流程代码或指南已发布

3端到端训练流程代码或指南已发布，且允许商用

评分方法

每个组成部分根据最符合的开放性类别，按0 至 3 分的定性标准评分。评估依据是所有公开的一手信息。

我们将这些因素综合为统一指标：Artificial Analysis Openness Index。计算方式如下：

对预训练和后训练的数据项目取平均值（数据部分共 6 分）
将所有组成部分的得分相加（最高 18/18 分）
将得分归一化到 0 至 100 分

如果模型基于第三方基础模型，其评分可能受到上游模型许可或信息披露范围的限制。对于增量版本或更新版本，我们只考虑明确针对新版本披露的信息（模型开发者可说明哪些部分与早期版本保持一致）。

Openness Index 排行榜

	开发者	模型	Openness Index	Intelligence Index	模型可用性	模型透明度	预训练数据访问	预训练数据许可	后训练数据访问	后训练数据许可
1	Institute of Foundation Models	K2 Think V2	88.89	11.50	6.00	10.00	3.00	1.00	3.00	1.00
2	Institute of Foundation Models	K2-V2 (High)	88.89	9.87	6.00	10.00	3.00	1.00	3.00	1.00
3	Institute of Foundation Models	K2-V2 (Medium)	88.89	9.02	6.00	10.00	3.00	1.00	3.00	1.00
4	Institute of Foundation Models	K2-V2 (Low)	88.89	7.31	6.00	10.00	3.00	1.00	3.00	1.00
5	Allen Institute for AI	Olmo 3.1 32B Think	88.89	7.12	6.00	10.00	3.00	1.00	3.00	1.00
6	Allen Institute for AI	Olmo 3.1 32B Instruct	88.89	6.50	6.00	10.00	3.00	1.00	3.00	1.00
7	Allen Institute for AI	Olmo 3 32B Think	88.89	6.47	6.00	10.00	3.00	1.00	3.00	1.00
8	Allen Institute for AI	OLMo 2 32B	88.89	5.97	6.00	10.00	3.00	1.00	3.00	1.00
9	Allen Institute for AI	Olmo 3 7B Think	88.89	5.62	6.00	10.00	3.00	1.00	3.00	1.00
10	Allen Institute for AI	OLMo 2 7B	88.89	5.58	6.00	10.00	3.00	1.00	3.00	1.00
11	Allen Institute for AI	Molmo 7B-D	88.89	5.56	6.00	10.00	3.00	1.00	3.00	1.00
12	Allen Institute for AI	Olmo 3 7B Instruct	88.89	5.24	6.00	10.00	3.00	1.00	3.00	1.00
13	Swiss AI Initiative	Apertus 70B Instruct	88.89	5.11	6.00	10.00	3.00	1.00	3.00	1.00
14	Swiss AI Initiative	Apertus 8B Instruct	88.89	4.83	6.00	10.00	3.00	1.00	3.00	1.00
15	NVIDIA	Nemotron 3.5 Super VL	83.33	26.49	6.00	9.00	2.00	1.00	2.00	1.00
16	NVIDIA	Nemotron 3 Ultra 550B A55B (Reasoning)	83.33	22.93	6.00	9.00	2.00	1.00	2.00	1.00
17	NVIDIA	Nemotron 3.5 Lightning	83.33	12.86	6.00	9.00	2.00	1.00	2.00	1.00
18	NVIDIA	Nemotron 3 Super 120B A12B (Reasoning)	83.33	12.83	6.00	9.00	2.00	1.00	2.00	1.00
19	NVIDIA	Nemotron Cascade 2 30B A3B	83.33	11.67	6.00	9.00	2.00	1.00	2.00	1.00
20	NVIDIA	NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)	83.33	8.90	6.00	9.00	2.00	1.00	2.00	1.00
21	OpenBMB	MiniCPM5-1B (Reasoning)	83.33	8.80	6.00	9.00	3.00	3.00	3.00	3.00
22	OpenBMB	MiniCPM5-1B (Non-reasoning)	83.33	8.69	6.00	9.00	3.00	3.00	3.00	3.00
23	NVIDIA	NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)	83.33	6.85	6.00	9.00	2.00	1.00	2.00	1.00
24	NVIDIA	NVIDIA Nemotron 3 Nano 4B	83.33	4.77	6.00	9.00	2.00	1.00	2.00	1.00
25	NVIDIA	Nemotron 3.5 Bolt 74B-A9B	72.22	21.41	6.00	7.00	2.00	1.00	2.00	1.00
26	IBM	Granite 4.2 30B	72.22	12.68	6.00	7.00	2.00	1.00	2.00	1.00
27	IBM	Granite 4.2 8B	72.22	11.08	6.00	7.00	2.00	1.00	2.00	1.00
28	Prime Intellect	INTELLECT-3 (Based on GLM-4.5-Air)	72.22	10.61	6.00	7.00	0.00	0.00	2.00	2.00
29	IBM	Granite 4.2 3B	72.22	9.06	6.00	7.00	2.00	1.00	2.00	1.00
30	NVIDIA	NVIDIA Nemotron Nano 12B v2 VL (Reasoning)	72.22	7.48	6.00	7.00	2.00	1.00	2.00	1.00
31	NVIDIA	NVIDIA Nemotron Nano 9B V2 (Reasoning)	72.22	7.43	6.00	7.00	2.00	1.00	2.00	1.00
32	NVIDIA	NVIDIA Nemotron Nano 9B V2 (Non-reasoning)	72.22	6.84	6.00	7.00	2.00	1.00	2.00	1.00
33	NVIDIA	NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)	72.22	5.82	6.00	7.00	2.00	1.00	2.00	1.00
34	Allen Institute for AI	Molmo2-8B	72.22	5.00	6.00	7.00	3.00	1.00	3.00	1.00
35	OpenBMB	MiniCPM5-2B	66.67	12.46	6.00	6.00	2.00	1.00	2.00	1.00
36	IBM	Granite 4.1 30B	61.11	7.44	6.00	5.00	2.00	1.00	2.00	1.00
37	Kimi	Kimi Linear 48B A3B Instruct	61.11	7.30	6.00	5.00	1.00	0.00	1.00	0.00
38	IBM	Granite 4.1 8B	61.11	6.57	6.00	5.00	2.00	1.00	2.00	1.00
39	IBM	Granite 4.1 3B	61.11	5.86	6.00	5.00	2.00	1.00	2.00	1.00
40	Z AI	GLM-4.5 (Reasoning)	55.56	12.77	6.00	4.00	0.00	0.00	0.00	0.00
41	Z AI	GLM-4.5-Air	55.56	11.09	6.00	4.00	0.00	0.00	0.00	0.00
42	NVIDIA	Nemotron 3 Nano Omni 30B A3B Reasoning	55.56	10.25	6.00	4.00	1.00	0.00	1.00	0.00
43	Baidu	ERNIE 4.5 300B A47B	55.56	7.50	6.00	4.00	0.00	0.00	0.00	0.00
44	IBM	Granite 4.0 H Small	55.56	6.05	6.00	4.00	2.00	1.00	2.00	1.00
45	IBM	Granite 4.0 H 1B	55.56	5.19	6.00	4.00	2.00	1.00	2.00	1.00
46	IBM	Granite 4.0 Micro	55.56	5.11	6.00	4.00	2.00	1.00	2.00	1.00
47	IBM	Granite 4.0 1B	55.56	5.01	6.00	4.00	2.00	1.00	2.00	1.00
48	IBM	Granite 4.0 350M	55.56	4.83	6.00	4.00	2.00	1.00	2.00	1.00
49	IBM	Granite 4.0 H 350M	55.56	4.83	6.00	4.00	2.00	1.00	2.00	1.00
50	Xiaomi	MiMo-V2.6-Pro	52.78	46.32	6.00	3.50	0.00	0.00	1.00	0.00
51	Xiaomi	MiMo-V2.6-Flash	52.78	37.88	6.00	3.50	0.00	0.00	1.00	0.00
52	Xiaomi	MiMo-V2-Flash (Reasoning)	52.78	20.82	6.00	3.50	0.00	0.00	1.00	0.00
53	Xiaomi	MiMo-V2-Flash (Non-reasoning)	52.78	16.05	6.00	3.50	0.00	0.00	1.00	0.00
54	NVIDIA	Llama Nemotron Super 49B v1.5 (Reasoning)	52.78	9.01	4.00	5.50	1.00	0.00	1.00	1.00
55	NVIDIA	Llama 3.3 Nemotron Super 49B v1 (Reasoning)	52.78	8.93	4.00	5.50	1.00	0.00	1.00	1.00
56	Z AI	GLM-4.5V (Reasoning)	52.78	7.56	6.00	3.50	1.00	0.00	0.00	0.00
57	NVIDIA	Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)	52.78	7.53	4.00	5.50	1.00	0.00	1.00	1.00
58	NVIDIA	Llama Nemotron Super 49B v1.5 (Non-reasoning)	52.78	7.36	4.00	5.50	1.00	0.00	1.00	1.00
59	NVIDIA	Llama 3.1 Nemotron Nano 4B v1.1 (Reasoning)	52.78	7.31	4.00	5.50	1.00	0.00	1.00	1.00
60	NVIDIA	Llama 3.3 Nemotron Super 49B v1 (Non-reasoning)	52.78	7.28	4.00	5.50	1.00	0.00	1.00	1.00
61	Z AI	GLM-4.5V (Non-reasoning)	52.78	6.70	6.00	3.50	1.00	0.00	0.00	0.00
62	DeepSeek	DeepSeek V4 Pro 0424 (Max)	50.00	30.45	6.00	3.00	0.00	0.00	0.00	0.00
63	DeepSeek	DeepSeek V4 Pro 0424 (High)	50.00	30.11	6.00	3.00	0.00	0.00	0.00	0.00
64	Z AI	GLM-5 (Reasoning)	50.00	27.91	6.00	3.00	0.00	0.00	0.00	0.00
65	DeepSeek	DeepSeek V4 Flash 0420 (High)	50.00	24.38	6.00	3.00	0.00	0.00	0.00	0.00
66	DeepSeek	DeepSeek V4 Flash 0420 (Max)	50.00	24.17	6.00	3.00	0.00	0.00	0.00	0.00
67	Xiaomi	MiMo-V2-Flash (Feb 2026)	50.00	22.44	6.00	3.00	0.00	0.00	0.00	0.00
68	Z AI	GLM-5 (Non-reasoning)	50.00	21.78	6.00	3.00	0.00	0.00	0.00	0.00
69	SK Telecom	A.X-K2	50.00	21.46	6.00	3.00	1.00	0.00	1.00	0.00
70	DeepSeek	DeepSeek V4 Pro 0424 (Non-reasoning)	50.00	20.83	6.00	3.00	0.00	0.00	0.00	0.00
71	DeepSeek	DeepSeek V4 Flash 0420 (Non-reasoning)	50.00	18.88	6.00	3.00	0.00	0.00	0.00	0.00
72	StepFun	Step 3.5 Flash	50.00	16.62	6.00	3.00	0.00	0.00	0.00	0.00
73	Alibaba	Qwen3 VL 235B A22B (Reasoning)	50.00	13.44	6.00	3.00	1.00	0.00	1.00	0.00
74	ServiceNow	Apriel-v1.6-15B-Thinker	50.00	13.40	6.00	3.00	0.00	0.00	0.00	0.00
75	DeepSeek	DeepSeek R1 0528 (May '25)	50.00	13.12	6.00	3.00	0.00	0.00	0.00	0.00
76	Alibaba	Qwen3 VL 32B (Reasoning)	50.00	11.87	6.00	3.00	1.00	0.00	1.00	0.00
77	Alibaba	Qwen3 VL 235B A22B Instruct	50.00	9.94	6.00	3.00	1.00	0.00	1.00	0.00
78	Alibaba	Qwen3 VL 30B A3B (Reasoning)	50.00	9.45	6.00	3.00	1.00	0.00	1.00	0.00
79	Mistral	Magistral Small 1.2	50.00	8.59	6.00	3.00	0.00	0.00	1.00	1.00
80	Alibaba	Qwen3 VL 32B Instruct	50.00	8.39	6.00	3.00	1.00	0.00	1.00	0.00
81	Alibaba	Qwen3 VL 8B (Reasoning)	50.00	8.17	6.00	3.00	1.00	0.00	1.00	0.00
82	Alibaba	Qwen3 VL 30B A3B Instruct	50.00	7.93	6.00	3.00	1.00	0.00	1.00	0.00
83	Alibaba	Qwen3 VL 8B Instruct	50.00	7.26	6.00	3.00	1.00	0.00	1.00	0.00
84	Alibaba	Qwen3 VL 4B (Reasoning)	50.00	7.05	6.00	3.00	1.00	0.00	1.00	0.00
85	Microsoft	Phi-4 Mini Instruct	50.00	6.33	6.00	3.00	1.00	0.00	1.00	0.00
86	Microsoft	Phi-4	50.00	5.92	6.00	3.00	1.00	0.00	1.00	0.00
87	Microsoft	Phi-4 Multimodal Instruct	50.00	5.81	6.00	3.00	1.00	0.00	1.00	0.00
88	Alibaba	Qwen3 VL 4B Instruct	50.00	5.66	6.00	3.00	1.00	0.00	1.00	0.00
89	Google	Gemma 3 27B Instruct	50.00	4.85	6.00	3.00	0.00	0.00	0.00	0.00
90	Google	Gemma 3 4B Instruct	50.00	4.83	6.00	3.00	0.00	0.00	0.00	0.00
91	Google	Gemma 3n E4B Instruct	50.00	4.83	6.00	3.00	0.00	0.00	0.00	0.00
92	Google	Gemma 3n E2B Instruct	50.00	4.83	6.00	3.00	0.00	0.00	0.00	0.00
93	Google	Gemma 3 1B Instruct	50.00	4.83	6.00	3.00	0.00	0.00	0.00	0.00
94	Google	Gemma 3 12B Instruct	50.00	3.75	6.00	3.00	0.00	0.00	0.00	0.00
95	ServiceNow	Apriel-v1.5-15B-Thinker	47.22	13.82	6.00	2.50	0.00	0.00	1.00	0.00
96	DeepSeek	DeepSeek R1 0528 Qwen3 8B	47.22	8.08	6.00	2.50	0.00	0.00	1.00	0.00
97	Nous Research	Hermes 4 - Llama-3.1 70B (Reasoning)	47.22	7.91	4.00	4.50	1.00	0.00	2.00	0.00
98	Nous Research	Hermes 4 - Llama-3.1 405B (Reasoning)	47.22	7.49	4.00	4.50	1.00	0.00	2.00	0.00
99	Nous Research	Hermes 4 - Llama-3.1 405B (Non-reasoning)	47.22	7.41	4.00	4.50	1.00	0.00	2.00	0.00
100	Nous Research	Hermes 4 - Llama-3.1 70B (Non-reasoning)	47.22	6.66	4.00	4.50	1.00	0.00	2.00	0.00
101	DeepSeek	DeepSeek V4.1 Flash 1017 (Max)	44.44	45.10	6.00	2.00	0.00	0.00	0.00	0.00
102	Z AI	GLM-5.3-Flash	44.44	41.81	6.00	2.00	0.00	0.00	0.00	0.00
103	DeepSeek	DeepSeek V4.1 Flash (Max)	44.44	39.46	6.00	2.00	0.00	0.00	0.00	0.00
104	DeepSeek	DeepSeek V4 Pro 0813 (Max)	44.44	36.00	6.00	2.00	0.00	0.00	0.00	0.00
105	DeepSeek	DeepSeek V4 Flash 0731 (Max)	44.44	34.33	6.00	2.00	0.00	0.00	0.00	0.00
106	Z AI	GLM-5.2 (Max)	44.44	33.71	6.00	2.00	0.00	0.00	0.00	0.00
107	Motif Technologies	Motif 3	44.44	33.57	6.00	2.00	0.00	0.00	0.00	0.00
108	Z AI	GLM-5.1 (Reasoning)	44.44	26.06	6.00	2.00	0.00	0.00	0.00	0.00
109	Tencent	Hy3	44.44	25.30	5.00	3.00	0.00	0.00	0.00	0.00
110	DeepSeek	DeepSeek V4.1 Flash (Non-reasoning)	44.44	24.67	6.00	2.00	0.00	0.00	0.00	0.00
111	InclusionAI	Ling-3.0-flash-VL	44.44	24.57	6.00	2.00	0.00	0.00	0.00	0.00
112	Z AI	GLM-5.1 (Non-reasoning)	44.44	24.24	6.00	2.00	0.00	0.00	0.00	0.00
113	InclusionAI	Ling-3.0-flash-Fin	44.44	22.60	6.00	2.00	0.00	0.00	0.00	0.00
114	Z AI	GLM-5.2 (Non-reasoning)	44.44	22.43	6.00	2.00	0.00	0.00	0.00	0.00
115	Z AI	GLM-4.7 (Reasoning)	44.44	22.24	6.00	2.00	0.00	0.00	0.00	0.00
116	DeepSeek	DeepSeek V4 Pro 0813 (Non-reasoning)	44.44	20.36	6.00	2.00	0.00	0.00	0.00	0.00
117	InclusionAI	Ling 3.0 Flash	44.44	20.13	6.00	2.00	0.00	0.00	0.00	0.00
118	LG AI Research	K-EXAONE 2.0 0803	44.44	19.69	6.00	2.00	0.00	0.00	0.00	0.00
119	Z AI	GLM-4.6 (Reasoning)	44.44	18.53	6.00	2.00	0.00	0.00	0.00	0.00
120	Meta	Muse Glimmer (High)	44.44	17.48	6.00	2.00	0.00	0.00	0.00	0.00
121	Z AI	GLM-4.7 (Non-reasoning)	44.44	17.36	6.00	2.00	0.00	0.00	0.00	0.00
122	DeepSeek	DeepSeek V3.2 Exp (Reasoning)	44.44	16.58	6.00	2.00	0.00	0.00	0.00	0.00
123	Z AI	GLM-4.6 (Non-reasoning)	44.44	14.93	6.00	2.00	0.00	0.00	0.00	0.00
124	Z AI	GLM-4.7-Flash (Reasoning)	44.44	14.87	6.00	2.00	0.00	0.00	0.00	0.00
125	DeepSeek	DeepSeek V3.2 Exp (Non-reasoning)	44.44	13.88	6.00	2.00	0.00	0.00	0.00	0.00
126	Kimi	Kimi K2	44.44	12.72	4.00	4.00	1.00	0.00	1.00	0.00
127	Alibaba	Qwen3 235B A22B 2507 (Reasoning)	44.44	12.71	6.00	2.00	0.00	0.00	0.00	0.00
128	ByteDance Seed	Seed-OSS-36B-Instruct	44.44	12.10	6.00	2.00	0.00	0.00	0.00	0.00
129	Alibaba	Qwen3 235B A22B 2507 Instruct	44.44	12.00	6.00	2.00	0.00	0.00	0.00	0.00
130	Alibaba	Qwen3 Coder 480B A35B Instruct	44.44	11.90	6.00	2.00	0.00	0.00	0.00	0.00
131	LongCat	LongCat Flash Lite	44.44	11.47	6.00	2.00	0.00	0.00	0.00	0.00
132	Alibaba	Qwen3 Next 80B A3B (Reasoning)	44.44	11.20	6.00	2.00	0.00	0.00	0.00	0.00
133	Arcee AI	Trinity Large Thinking	44.44	10.82	6.00	2.00	0.00	0.00	0.00	0.00
134	Z AI	GLM-4.7-Flash (Non-reasoning)	44.44	10.56	6.00	2.00	0.00	0.00	0.00	0.00
135	Alibaba	Qwen3 30B A3B 2507 (Reasoning)	44.44	9.81	6.00	2.00	0.00	0.00	0.00	0.00
136	Alibaba	Qwen3 Next 80B A3B Instruct	44.44	9.64	6.00	2.00	0.00	0.00	0.00	0.00
137	Trillion Labs	Tri-21B-think Preview	44.44	9.59	6.00	2.00	0.00	0.00	0.00	0.00
138	Alibaba	Qwen3 Coder 30B A3B Instruct	44.44	9.59	6.00	2.00	0.00	0.00	0.00	0.00
139	Google	DiffusionGemma 26B A4B	44.44	9.51	6.00	2.00	0.00	0.00	0.00	0.00
140	InclusionAI	Ling-1T	44.44	9.17	6.00	2.00	0.00	0.00	0.00	0.00
141	Alibaba	Qwen3 4B 2507 (Reasoning)	44.44	8.80	6.00	2.00	0.00	0.00	0.00	0.00
142	TII UAE	Falcon-H1R-7B	44.44	7.83	4.00	4.00	1.00	0.00	1.00	0.00
143	InclusionAI	Ling-flash-2.0	44.44	7.81	6.00	2.00	0.00	0.00	0.00	0.00
144	Alibaba	Qwen3 Omni 30B A3B (Reasoning)	44.44	7.76	6.00	2.00	0.00	0.00	0.00	0.00
145	Mistral	Devstral Small (Jul '25)	44.44	7.60	6.00	2.00	0.00	0.00	0.00	0.00
146	Alibaba	Qwen3 30B A3B 2507 Instruct	44.44	7.52	6.00	2.00	0.00	0.00	0.00	0.00
147	NVIDIA	Llama 3.1 Nemotron Instruct 70B	44.44	6.94	4.00	4.00	0.00	0.00	1.00	1.00
148	Alibaba	Qwen3 4B 2507 Instruct	44.44	6.74	6.00	2.00	0.00	0.00	0.00	0.00
149	Alibaba	Qwen3 Omni 30B A3B Instruct	44.44	6.01	6.00	2.00	0.00	0.00	0.00	0.00
150	OpenBMB	MiniCPM-V 4.6 1.3B	44.44	5.69	6.00	2.00	0.00	0.00	0.00	0.00
151	InclusionAI	Ling-mini-2.0	44.44	5.54	6.00	2.00	0.00	0.00	0.00	0.00
152	Google	Gemma 3 270M	44.44	5.12	6.00	2.00	0.00	0.00	0.00	0.00
153	Cohere	North Mini Code	41.67	9.91	6.00	1.50	0.00	0.00	1.00	0.00
154	Alibaba	Qwen3 Coder Next	41.67	9.24	6.00	1.50	0.00	0.00	1.00	0.00
155	Nanbeige	Nanbeige4.1-3B	41.67	8.40	6.00	1.50	0.00	0.00	1.00	0.00
156	StepFun	Step3 VL 10B	41.67	7.69	6.00	1.50	0.00	0.00	1.00	0.00
157	Kimi	Kimi K3 (Max)	38.89	43.59	4.00	3.00	0.00	0.00	0.00	0.00
158	InclusionAI	Ling 3.1 Flash	38.89	41.09	6.00	1.00	0.00	0.00	0.00	0.00
159	Alibaba	Qwen3.8-Flash-Next	38.89	39.82	5.00	2.00	0.00	0.00	0.00	0.00
160	Alibaba	Qwen3.8 27B (Xhigh)	38.89	33.68	6.00	1.00	0.00	0.00	0.00	0.00
161	Reflection	Beam 1.0	38.89	30.33	6.00	1.00	0.00	0.00	0.00	0.00
162	Kimi	Kimi K3 (Low)	38.89	30.07	4.00	3.00	0.00	0.00	0.00	0.00
163	Nex AGI	Nex-N2-Pro (Based on Qwen3.5-397B-A17B)	38.89	28.20	6.00	1.00	0.00	0.00	0.00	0.00
164	Alibaba	Qwen3.8 27B (Medium)	38.89	27.55	6.00	1.00	0.00	0.00	0.00	0.00
165	Alibaba	Qwen3.8 27B (Low)	38.89	26.20	6.00	1.00	0.00	0.00	0.00	0.00
166	Xiaomi	MiMo-V2.5-Pro (Reasoning)	38.89	25.99	6.00	1.00	0.00	0.00	0.00	0.00
167	Thinking Machines	Inkling (Xhigh)	38.89	24.98	6.00	1.00	0.00	0.00	0.00	0.00
168	Alibaba	Qwen3.5 27B (Reasoning)	38.89	22.90	6.00	1.00	0.00	0.00	0.00	0.00
169	AI9Stars	G9v3-39A5B	38.89	21.83	6.00	1.00	0.00	0.00	0.00	0.00
170	Xiaomi	MiMo-V2.5	38.89	21.82	6.00	1.00	0.00	0.00	0.00	0.00
171	Alibaba	Qwen3.5 397B A17B (Non-reasoning)	38.89	21.44	6.00	1.00	0.00	0.00	0.00	0.00
172	Alibaba	Qwen3.6 27B (Reasoning)	38.89	21.43	6.00	1.00	0.00	0.00	0.00	0.00
173	Alibaba	Qwen3.8 27B (Non-reasoning)	38.89	20.15	6.00	1.00	0.00	0.00	0.00	0.00
174	Alibaba	Qwen3.6 27B (Non-reasoning)	38.89	19.83	6.00	1.00	0.00	0.00	0.00	0.00
175	StepFun	Step 3.7 Flash	38.89	19.48	6.00	1.00	0.00	0.00	0.00	0.00
176	Alibaba	Qwen3.5 27B (Non-reasoning)	38.89	19.37	6.00	1.00	0.00	0.00	0.00	0.00
177	Alibaba	Qwen3.5 35B A3B (Reasoning)	38.89	19.33	6.00	1.00	0.00	0.00	0.00	0.00
178	LongCat	LongCat 2.0	38.89	19.11	6.00	1.00	0.00	0.00	0.00	0.00
179	Alibaba	Qwen3.5 397B A17B (Reasoning)	38.89	18.42	6.00	1.00	0.00	0.00	0.00	0.00
180	Xiaomi	MiMo-V2.5-Pro (Non-reasoning)	38.89	18.29	6.00	1.00	0.00	0.00	0.00	0.00
181	Alibaba	Qwen3.6 35B A3B (Reasoning)	38.89	18.23	6.00	1.00	0.00	0.00	0.00	0.00
182	Alibaba	Qwen3.5 122B A10B (Non-reasoning)	38.89	17.75	6.00	1.00	0.00	0.00	0.00	0.00
183	InclusionAI	Ling-2.6-1T	38.89	17.00	6.00	1.00	0.00	0.00	0.00	0.00
184	Google	Gemma 4 26B A4B (Reasoning)	38.89	16.67	6.00	1.00	0.00	0.00	0.00	0.00
185	InclusionAI	Ring-2.6-1T	38.89	16.62	6.00	1.00	0.00	0.00	0.00	0.00
186	Alibaba	Qwen3.5 122B A10B (Reasoning)	38.89	15.57	6.00	1.00	0.00	0.00	0.00	0.00
187	Alibaba	Qwen3.6 35B A3B (Non-reasoning)	38.89	15.23	6.00	1.00	0.00	0.00	0.00	0.00
188	Alibaba	Qwen3.5 35B A3B (Non-reasoning)	38.89	15.12	6.00	1.00	0.00	0.00	0.00	0.00
189	DeepSeek	DeepSeek V3.1 Terminus (Reasoning)	38.89	14.76	6.00	1.00	0.00	0.00	0.00	0.00
190	Google	Gemma 4 31B (Reasoning)	38.89	14.67	6.00	1.00	0.00	0.00	0.00	0.00
191	Google	Gemma 4 12B (Reasoning)	38.89	14.18	6.00	1.00	0.00	0.00	0.00	0.00
192	DeepSeek	DeepSeek V3.1 Terminus (Non-reasoning)	38.89	13.93	6.00	1.00	0.00	0.00	0.00	0.00
193	Google	Gemma 4 31B (Non-reasoning)	38.89	13.92	6.00	1.00	0.00	0.00	0.00	0.00
194	Alibaba	Qwen3.5 9B (Non-reasoning)	38.89	13.27	6.00	1.00	0.00	0.00	0.00	0.00
195	Cohere	Command A+	38.89	13.13	6.00	1.00	0.00	0.00	0.00	0.00
196	Google	Gemma 4 26B A4B (Non-reasoning)	38.89	13.13	6.00	1.00	0.00	0.00	0.00	0.00
197	Alibaba	Qwen3.5 4B (Reasoning)	38.89	13.12	6.00	1.00	0.00	0.00	0.00	0.00
198	Multiverse Computing	HyperNova 60B 2605 (High, Based on gpt-oss-120b)	38.89	11.74	6.00	1.00	0.00	0.00	0.00	0.00
199	OpenAI	gpt-oss-120b (High)	38.89	11.60	6.00	1.00	0.00	0.00	0.00	0.00
200	Mistral	Mistral Small 4 (Reasoning)	38.89	11.27	6.00	1.00	0.00	0.00	0.00	0.00
201	Alibaba	Qwen3.5 9B (Reasoning)	38.89	11.20	6.00	1.00	0.00	0.00	0.00	0.00
202	InclusionAI	Ling 3.0 Tiny	38.89	11.06	6.00	1.00	0.00	0.00	0.00	0.00
203	InclusionAI	Ring-1T	38.89	10.90	6.00	1.00	0.00	0.00	0.00	0.00
204	Alibaba	Qwen3.5 4B (Non-reasoning)	38.89	10.81	6.00	1.00	0.00	0.00	0.00	0.00
205	OpenAI	gpt-oss-120b (Low)	38.89	10.21	6.00	1.00	0.00	0.00	0.00	0.00
206	OpenAI	gpt-oss-20b (Low)	38.89	9.95	6.00	1.00	0.00	0.00	0.00	0.00
207	InclusionAI	Ling 2.6 Flash	38.89	9.68	6.00	1.00	0.00	0.00	0.00	0.00
208	Google	Gemma 4 12B (Non-reasoning)	38.89	9.38	6.00	1.00	0.00	0.00	0.00	0.00
209	Mistral	Mistral Large 3	38.89	9.27	6.00	1.00	0.00	0.00	0.00	0.00
210	Mistral	Mistral Small 4 (Non-reasoning)	38.89	8.99	6.00	1.00	0.00	0.00	0.00	0.00
211	Trillion Labs	Tri-21B-Think	38.89	8.99	6.00	1.00	0.00	0.00	0.00	0.00
212	OpenAI	gpt-oss-20b (High)	38.89	8.97	6.00	1.00	0.00	0.00	0.00	0.00
213	Google	Gemma 4 E4B (Reasoning)	38.89	8.91	6.00	1.00	0.00	0.00	0.00	0.00
214	Sarvam	Sarvam 105B (High)	38.89	8.79	6.00	1.00	0.00	0.00	0.00	0.00
215	Liquid AI	LFM2.5-2.6B	38.89	8.39	5.00	2.00	0.00	0.00	0.00	0.00
216	Google	Gemma 4 E2B (Reasoning)	38.89	7.77	6.00	1.00	0.00	0.00	0.00	0.00
217	Meta	Llama 3.3 Instruct 70B	38.89	7.66	4.00	3.00	1.00	0.00	1.00	0.00
218	Google	Gemma 4 E4B (Non-reasoning)	38.89	7.46	6.00	1.00	0.00	0.00	0.00	0.00
219	Meta	Llama 3.1 Instruct 405B	38.89	7.29	4.00	3.00	1.00	0.00	1.00	0.00
220	InclusionAI	Ring-flash-2.0	38.89	7.15	6.00	1.00	0.00	0.00	0.00	0.00
221	Mistral	Mistral Small 3.2	38.89	6.96	6.00	1.00	0.00	0.00	0.00	0.00
222	Alibaba	Qwen3.5 2B (Reasoning)	38.89	6.94	6.00	1.00	0.00	0.00	0.00	0.00
223	Sarvam	Sarvam 30B (High)	38.89	6.56	6.00	1.00	0.00	0.00	0.00	0.00
224	Google	Gemma 4 E2B (Non-reasoning)	38.89	6.48	6.00	1.00	0.00	0.00	0.00	0.00
225	Perplexity	R1 1776	38.89	6.44	6.00	1.00	0.00	0.00	0.00	0.00
226	Meta	Llama 3.2 Instruct 90B (Vision)	38.89	6.41	4.00	3.00	1.00	0.00	1.00	0.00
227	Alibaba	Qwen3.5 2B (Non-reasoning)	38.89	6.19	6.00	1.00	0.00	0.00	0.00	0.00
228	Alibaba	Qwen3.5 0.8B (Reasoning)	38.89	6.13	6.00	1.00	0.00	0.00	0.00	0.00
229	Nous Research	DeepHermes 3 - Mistral 24B Preview (Non-reasoning)	38.89	6.07	6.00	1.00	0.00	0.00	0.00	0.00
230	Mistral	Ministral 3 14B	38.89	6.04	6.00	1.00	0.00	0.00	0.00	0.00
231	AI21 Labs	Jamba Reasoning 3B	38.89	5.67	6.00	1.00	0.00	0.00	0.00	0.00
232	Reka AI	Reka Flash 3	38.89	5.65	6.00	1.00	0.00	0.00	0.00	0.00
233	Mistral	Ministral 3 8B	38.89	5.48	6.00	1.00	0.00	0.00	0.00	0.00
234	Meta	Llama 3.2 Instruct 11B (Vision)	38.89	5.41	4.00	3.00	1.00	0.00	1.00	0.00
235	Alibaba	Qwen3.5 0.8B (Non-reasoning)	38.89	5.40	6.00	1.00	0.00	0.00	0.00	0.00
236	Mistral	Ministral 3 3B	38.89	4.84	6.00	1.00	0.00	0.00	0.00	0.00
237	Deep Cogito	Cogito v2.1 (Reasoning)	38.89	-	6.00	1.00	0.00	0.00	0.00	0.00
238	DeepSeek	DeepSeek R1 Distill Llama 70B	36.11	7.89	4.00	2.50	0.00	0.00	1.00	0.00
239	Cohere	Tiny Aya Global	36.11	4.83	3.00	3.50	0.00	0.00	1.00	0.00
240	Z AI	GLM-5.3 (Max)	33.33	44.78	5.00	1.00	0.00	0.00	0.00	0.00
241	StepFun	Step 5 Preview	33.33	43.73	4.00	2.00	0.00	0.00	0.00	0.00
242	Z AI	GLM-5.3 (Low)	33.33	34.30	5.00	1.00	0.00	0.00	0.00	0.00
243	MiniMax	MiniMax-M3	33.33	29.22	4.00	2.00	0.00	0.00	0.00	0.00
244	Kimi	Kimi K2.6 (Reasoning)	33.33	26.98	4.00	2.00	0.00	0.00	0.00	0.00
245	Thinking Machines	Inkling Small	33.33	25.66	5.00	1.00	0.00	0.00	0.00	0.00
246	Upstage	Solar Open2 250B	33.33	24.74	4.00	2.00	0.00	0.00	0.00	0.00
247	Kimi	Kimi K2.6 (Non-reasoning)	33.33	23.57	4.00	2.00	0.00	0.00	0.00	0.00
248	Kimi	Kimi K2.5 (Reasoning)	33.33	23.46	4.00	2.00	0.00	0.00	0.00	0.00
249	Tencent	Hy3-preview (Reasoning)	33.33	22.74	5.00	1.00	0.00	0.00	0.00	0.00
250	Kimi	Kimi K2.5 (Non-reasoning)	33.33	19.43	4.00	2.00	0.00	0.00	0.00	0.00
251	Tencent	Hy3-preview (Non-reasoning)	33.33	17.04	5.00	1.00	0.00	0.00	0.00	0.00
252	Mistral	Mistral Medium 3.5	33.33	14.19	5.00	1.00	0.00	0.00	0.00	0.00
253	AI9Stars	G9v3-3B	33.33	10.84	6.00	0.00	0.00	0.00	0.00	0.00
254	Liquid AI	LFM2.5-8B-A1B	33.33	7.22	4.00	2.00	0.00	0.00	0.00	0.00
255	Cohere	Command A	33.33	6.96	3.00	3.00	0.00	0.00	0.00	0.00
256	Liquid AI	LFM2 2.6B	33.33	5.21	4.00	2.00	0.00	0.00	0.00	0.00
257	Nous Research	DeepHermes 3 - Llama-3.1 8B Preview (Non-reasoning)	33.33	5.08	5.00	1.00	0.00	0.00	0.00	0.00
258	Liquid AI	LFM2 8B A1B	33.33	4.93	4.00	2.00	0.00	0.00	0.00	0.00
259	Liquid AI	LFM2 1.2B	33.33	4.83	4.00	2.00	0.00	0.00	0.00	0.00
260	Naver	HyperCLOVA X SEED Think (32B)	30.56	11.36	4.00	1.50	1.00	0.00	0.00	0.00
261	Alibaba	Qwen3.8 2.4T A95B	27.78	39.95	4.00	1.00	0.00	0.00	0.00	0.00
262	Meta	Muse Spark Open (Xhigh)	27.78	38.87	4.00	1.00	0.00	0.00	0.00	0.00
263	Kimi	Kimi K2.7 Code	27.78	25.81	4.00	1.00	0.00	0.00	0.00	0.00
264	MiniMax	MiniMax-M2.5	27.78	22.80	4.00	1.00	0.00	0.00	0.00	0.00
265	Kimi	Kimi K2 Thinking	27.78	22.02	4.00	1.00	0.00	0.00	0.00	0.00
266	MiniMax	MiniMax-M2.1	27.78	20.95	4.00	1.00	0.00	0.00	0.00	0.00
267	MiniMax	MiniMax-M2	27.78	18.63	4.00	1.00	0.00	0.00	0.00	0.00
268	Kimi	Kimi K2 0905	27.78	15.29	4.00	1.00	0.00	0.00	0.00	0.00
269	LG AI Research	K-EXAONE (Reasoning)	27.78	14.37	4.00	1.00	0.00	0.00	0.00	0.00
270	LG AI Research	EXAONE 4.5 33B (Reasoning)	27.78	13.21	3.00	2.00	0.00	0.00	0.00	0.00
271	Mistral	Magistral Medium 1.2	27.78	11.83	2.00	3.00	0.00	0.00	1.00	1.00
272	LG AI Research	K-EXAONE (Non-reasoning)	27.78	11.21	4.00	1.00	0.00	0.00	0.00	0.00
273	Upstage	Solar Open 100B (Reasoning)	27.78	10.36	4.00	1.00	0.00	0.00	0.00	0.00
274	Meta	Llama 4 Maverick	27.78	9.30	4.00	1.00	0.00	0.00	0.00	0.00
275	Mistral	Devstral 2	27.78	8.62	5.00	0.00	0.00	0.00	0.00	0.00
276	LG AI Research	EXAONE 4.0 32B (Reasoning)	27.78	8.18	3.00	2.00	0.00	0.00	0.00	0.00
277	Mistral	Devstral Small 2	27.78	7.54	5.00	0.00	0.00	0.00	0.00	0.00
278	Meta	Llama 4 Scout	27.78	6.45	4.00	1.00	0.00	0.00	0.00	0.00
279	LG AI Research	EXAONE 4.0 32B (Non-reasoning)	27.78	6.33	3.00	2.00	0.00	0.00	0.00	0.00
280	Liquid AI	LFM2 24B A2B	27.78	5.95	4.00	1.00	0.00	0.00	0.00	0.00
281	LG AI Research	Exaone 4.0 1.2B (Reasoning)	27.78	5.27	3.00	2.00	0.00	0.00	0.00	0.00
282	LG AI Research	Exaone 4.0 1.2B (Non-reasoning)	27.78	5.23	3.00	2.00	0.00	0.00	0.00	0.00
283	Liquid AI	LFM2.5-1.2B-Thinking	27.78	5.22	4.00	1.00	0.00	0.00	0.00	0.00
284	Liquid AI	LFM2.5-1.2B-Instruct	27.78	5.21	4.00	1.00	0.00	0.00	0.00	0.00
285	Liquid AI	LFM2.5-VL-1.6B	27.78	4.83	4.00	1.00	0.00	0.00	0.00	0.00
286	LG AI Research	EXAONE 4.5 33B (Non-reasoning)	27.78	-	3.00	2.00	0.00	0.00	0.00	0.00
287	MiniMax	MiniMax-M2.7	22.22	22.76	3.00	1.00	0.00	0.00	0.00	0.00
288	AI21 Labs	Jamba 1.7 Large	22.22	6.07	4.00	0.00	0.00	0.00	0.00	0.00
289	AI21 Labs	Jamba 1.7 Mini	22.22	5.22	4.00	0.00	0.00	0.00	0.00	0.00
290	Alibaba	Qwen3 Max Thinking (Preview)	16.67	16.29	2.00	1.00	0.00	0.00	0.00	0.00
291	Alibaba	Qwen3 Max	16.67	15.61	2.00	1.00	0.00	0.00	0.00	0.00
292	Anthropic	Claude Opus 4.5 (Reasoning)	11.11	29.10	2.00	0.00	0.00	0.00	0.00	0.00
293	Anthropic	Claude Opus 4.5 (Non-reasoning)	11.11	23.68	2.00	0.00	0.00	0.00	0.00	0.00
294	Anthropic	Claude 4.5 Sonnet (Reasoning)	11.11	20.67	2.00	0.00	0.00	0.00	0.00	0.00
295	Anthropic	Claude 4.5 Sonnet (Non-reasoning)	11.11	19.34	2.00	0.00	0.00	0.00	0.00	0.00
296	ByteDance Seed	Doubao Seed Code	11.11	16.95	2.00	0.00	0.00	0.00	0.00	0.00
297	Anthropic	Claude 4.5 Haiku (Reasoning)	11.11	16.88	2.00	0.00	0.00	0.00	0.00	0.00
298	Anthropic	Claude 4.5 Haiku (Non-reasoning)	11.11	15.41	2.00	0.00	0.00	0.00	0.00	0.00
299	SpaceXAI	Grok 3 mini Reasoning (High)	11.11	14.62	2.00	0.00	0.00	0.00	0.00	0.00
300	OpenAI	GPT-5 (ChatGPT)	11.11	10.44	2.00	0.00	0.00	0.00	0.00	0.00
301	Mistral	Mistral Medium 3.1	11.11	9.18	2.00	0.00	0.00	0.00	0.00	0.00
302	Amazon	Nova Premier	11.11	9.16	2.00	0.00	0.00	0.00	0.00	0.00
303	Mistral	Devstral Medium	11.11	9.01	2.00	0.00	0.00	0.00	0.00	0.00
304	Upstage	Solar Pro 2 (Reasoning)	11.11	7.49	2.00	0.00	0.00	0.00	0.00	0.00
305	Upstage	Solar Pro 2 (Non-reasoning)	11.11	7.00	2.00	0.00	0.00	0.00	0.00	0.00
306	Amazon	Nova Pro	11.11	6.96	2.00	0.00	0.00	0.00	0.00	0.00
307	Amazon	Nova Lite	11.11	6.67	2.00	0.00	0.00	0.00	0.00	0.00
308	Amazon	Nova Micro	11.11	5.88	2.00	0.00	0.00	0.00	0.00	0.00
309	Google	Gemini 3 Pro Preview (High)	5.56	27.96	1.00	0.00	0.00	0.00	0.00	0.00
310	OpenAI	GPT-5 Codex (High)	5.56	24.88	1.00	0.00	0.00	0.00	0.00	0.00
311	OpenAI	GPT-5.1 (High)	5.56	24.74	1.00	0.00	0.00	0.00	0.00	0.00
312	OpenAI	GPT-5 (High)	5.56	22.98	1.00	0.00	0.00	0.00	0.00	0.00
313	OpenAI	GPT-5 (Medium)	5.56	22.87	1.00	0.00	0.00	0.00	0.00	0.00
314	SpaceXAI	Grok 4	5.56	22.49	1.00	0.00	0.00	0.00	0.00	0.00
315	Google	Gemini 3 Pro Preview (Low)	5.56	22.32	1.00	0.00	0.00	0.00	0.00	0.00
316	OpenAI	GPT-5 (Low)	5.56	20.79	1.00	0.00	0.00	0.00	0.00	0.00
317	OpenAI	GPT-5 mini (Medium)	5.56	20.60	1.00	0.00	0.00	0.00	0.00	0.00
318	SpaceXAI	Grok 4.1 Fast (Reasoning)	5.56	20.37	1.00	0.00	0.00	0.00	0.00	0.00
319	OpenAI	o3	5.56	20.20	1.00	0.00	0.00	0.00	0.00	0.00
320	SpaceXAI	Grok 4 Fast (Reasoning)	5.56	17.94	1.00	0.00	0.00	0.00	0.00	0.00
321	OpenAI	GPT-5 mini (High)	5.56	16.77	1.00	0.00	0.00	0.00	0.00	0.00
322	Google	Gemini 2.5 Pro	5.56	16.08	1.00	0.00	0.00	0.00	0.00	0.00
323	Google	Gemini 2.5 Flash Preview (Sep '25) (Reasoning)	5.56	15.47	1.00	0.00	0.00	0.00	0.00	0.00
324	SpaceXAI	Grok Code Fast 1	5.56	14.06	1.00	0.00	0.00	0.00	0.00	0.00
325	OpenAI	GPT-5.1 (Non-reasoning)	5.56	13.31	1.00	0.00	0.00	0.00	0.00	0.00
326	OpenAI	GPT-5 nano (High)	5.56	12.99	1.00	0.00	0.00	0.00	0.00	0.00
327	OpenAI	GPT-5 nano (Medium)	5.56	12.48	1.00	0.00	0.00	0.00	0.00	0.00
328	Google	Gemini 2.5 Flash Preview (Sep '25) (Non-reasoning)	5.56	12.38	1.00	0.00	0.00	0.00	0.00	0.00
329	OpenAI	GPT-5 (Minimal)	5.56	11.45	1.00	0.00	0.00	0.00	0.00	0.00
330	SpaceXAI	Grok 4.1 Fast (Non-reasoning)	5.56	11.28	1.00	0.00	0.00	0.00	0.00	0.00
331	SpaceXAI	Grok 4 Fast (Non-reasoning)	5.56	11.06	1.00	0.00	0.00	0.00	0.00	0.00
332	Google	Gemini 2.5 Flash-Lite Preview (Sep '25) (Reasoning)	5.56	10.35	1.00	0.00	0.00	0.00	0.00	0.00
333	OpenAI	GPT-5 mini (Minimal)	5.56	9.91	1.00	0.00	0.00	0.00	0.00	0.00
334	Google	Gemini 2.5 Flash-Lite Preview (Sep '25) (Non-reasoning)	5.56	9.34	1.00	0.00	0.00	0.00	0.00	0.00
335	OpenAI	GPT-5 nano (Minimal)	5.56	7.09	1.00	0.00	0.00	0.00	0.00	0.00

常见问题

Artificial Analysis Openness Index 是一个综合分数，用于衡量 AI 模型的开放可用程度和文档透明度，包括模型权重、许可、训练数据、训练方法和训练代码。

Artificial Analysis Openness Index 在模型可用性（权重访问和许可）以及模型透明度（训练数据访问和许可、方法披露、训练流水线发布）方面，以 0 到 3 的定性等级为模型评分。数据分数取预训练和后训练阶段的平均值，各组成部分最高合计 18 分，再归一化为 0 到 100 分。

在已发布 Openness Index 结果的模型中，K2 Think V2 目前以 89 获得最高分。查看模型

开放透明的模型有助于复现、独立评估、针对具体场景进行定制，并避免供应商锁定地用于商业用途。Artificial Analysis Openness Index 以一个可比较的分数表示模型的开放程度，让用户在选择模型时可以同时权衡开放性、智能、速度和价格。

不等同。开放权重表示模型权重可用，但不保证源代码、训练数据、训练方法或宽松许可也可用。Artificial Analysis Openness Index 分别评估这些因素，以区分仅可访问的模型与广泛开放且透明的模型。

探索评测

Artificial Analysis Intelligence Index v4.3.2

A composite benchmark aggregating ten challenging evaluations to provide a holistic measure of AI capabilities across mathematics, science, coding, and reasoning.

AA-Briefcase v1.1: Agentic Knowledge Work Benchmark

A private evaluation developed by Artificial Analysis for frontier agentic capability in long-horizon knowledge work, testing agents on realistic business workflows that require deliverables such as spreadsheets, presentations, and memos.

GDPval-AA v2.1 Benchmark

GDPval-AA v2.1 is Artificial Analysis' evaluation framework for OpenAI's GDPval dataset. It tests AI models on real-world tasks across 44 occupations and 9 major industries. Models are given shell access and web browsing capabilities in an agentic loop via Stirrup to solve tasks, with Elo ratings derived from blind pairwise comparisons.

AA-3D: Interactive 3D Experience Benchmark

A benchmark developed by Artificial Analysis for how well AI models build interactive 3D experiences. Models work as agents in an offline sandbox via Stirrup to turn a written brief into a three.js scene, with Elo ratings derived from blind pairwise votes.

APEX-Agents-AA Benchmark

Artificial Analysis' implementation of the APEX-Agents benchmark, testing AI agents on long-horizon, cross-application tasks in professional-services environments with realistic application tooling.

AA-AnalystAgent Benchmark

Artificial Analysis' data analysis benchmark, testing AI agents on their ability to work with spreadsheets and documents to answer quantitative questions a Business Analyst or Data Analyst would face day-to-day.

AutomationBench-AA: Agentic SaaS Workflow Benchmark

A benchmark measuring agentic task completion across simulated SaaS application environments, scoring the share of each task's objectives completed without guardrail violations.

Harvey LAB-AA v1.1 Benchmark

Artificial Analysis' implementation of Harvey's Legal Agent Benchmark (LAB), testing AI agents on real-world legal work from Harvey's dataset of 120 private tasks spanning 24 legal practice areas. v1.1, built in collaboration with Harvey, checks every deliverable against the task's source documents for hallucinations, so work that passes every rubric criterion but contains a material hallucination receives no credit. A panel of three judges grades each deliverable criterion-by-criterion.

GDP.pdf Benchmark

Artificial Analysis' implementation of Surge AI's GDP.pdf benchmark, testing whether language models can reason over long, real-world professional documents and satisfy detailed task-specific criteria.

EnterpriseOps-Gym-AA Benchmark

Artificial Analysis' independent implementation of ServiceNow's EnterpriseOps-Gym, an agentic benchmark testing whether LLM agents can complete stateful, multi-step enterprise workflows across eight business domains via live tool use, graded on the final state of the underlying databases.

Terminal-Bench 4.0 Benchmark

A harder 66-task benchmark of complex terminal work across software, machine learning, science, operations, security, hardware, and media, with recalibrated compute and time allowances and improved instructions, environments, and verifiers.

Terminal-Bench-Science 0.1 Benchmark

A 70-task benchmark of research workflows authored and reviewed by domain experts across the life, physical, earth, mathematical, and engineering sciences, each completed in a terminal and checked by its own set of tests.

FrontierCode 1.1 Benchmark

Artificial Analysis' implementation of Cognition's FrontierCode 1.1, a private agentic coding benchmark that tests models on completing coding tasks in a way that would actually get merged by maintainers of real-world repositories.

AA-PersonalityBench: AI Personality & Communication Style Benchmark

Measures how models communicate rather than whether they are correct: 30 judged personality traits rolled up into 8 indexes, plus 21 measured communication characteristics.

AA-LCR v1.1: Long Context Reasoning Benchmark

A challenging benchmark measuring language models' ability to extract, reason about, and synthesize information from long-form documents ranging from 10k to 100k tokens (measured using the cl100k_base tokenizer).

AA-Omniscience: Knowledge and Hallucination Benchmark

A benchmark measuring factual recall and hallucination across various economically relevant domains.

SciCode Benchmark

A scientist-curated coding benchmark featuring 288 test set subproblems from 80 laboratory problems across 16 scientific disciplines.

Humanity's Last Exam Benchmark

A frontier-level benchmark with 2,500 expert-vetted questions across mathematics, sciences, and humanities, designed to be the final closed-ended academic evaluation.

CritPt Benchmark

A benchmark designed to test LLMs on research-level physics reasoning tasks, featuring 71 composite research challenges.

GPQA Diamond Benchmark

The most challenging 198 questions from GPQA, where PhD experts achieve 65% accuracy but skilled non-experts only reach 34% despite web access.

ITBench-AA Benchmark

Artificial Analysis' implementation of IBM's ITBench benchmark, testing AI agents on Kubernetes incident root-cause analysis from offline incident snapshots. The agent inspects alerts, events, traces, and topology and identifies the contributing-factor entities (deployments, pods, namespaces, network policies, etc.) responsible for the failure.

MMMU-Pro Benchmark

An enhanced MMMU benchmark that eliminates shortcuts and guessing strategies to more rigorously test multimodal models across 30 academic disciplines.

IFBench Benchmark

A benchmark evaluating precise instruction-following generalization on 58 diverse, verifiable out-of-domain constraints that test models' ability to follow specific output requirements.

MLCR-AA: Medical Long Context Reasoning Benchmark

An open benchmark from Wisedocs measuring how well models reason over long, fragmented medical records, performing the multi-document synthesis claims professionals rely on when reviewing insurance and healthcare cases.

HealthBench Benchmark

OpenAI's HealthBench: 5,000 health conversations graded against physician-written rubrics, from emergency referrals to global health.

HealthBench Professional Benchmark

OpenAI's HealthBench Professional: 525 clinician conversations covering consult, writing and research tasks, graded against physician-written rubrics.

𝜏³-Banking Benchmark

A fintech customer-support benchmark from the 𝜏-Knowledge framework that tests whether agents can navigate a large unstructured knowledge base and execute multi-step tool calls to resolve realistic banking workflows.

Terminal-Bench 2.1 Benchmark

A verified refresh of Terminal-Bench 2.0 — 89 curated tasks across software engineering, system administration, data processing, model training, and security, with environment and instruction fixes so scores reflect agent capability rather than environment gaps.

Terminal-Bench Hard Benchmark

An agentic benchmark evaluating AI capabilities in terminal environments through software engineering, system administration, and data processing tasks.

𝜏²-Bench Telecom Benchmark

A dual-control conversational AI benchmark simulating technical support scenarios where both agent and user must coordinate actions to resolve telecom service issues.

MMLU-Pro Benchmark

An enhanced version of MMLU with 12,000 graduate-level questions across 14 subject areas, featuring ten answer options and deeper reasoning requirements.

LiveCodeBench Benchmark

A contamination-free coding benchmark that continuously harvests fresh competitive programming problems from LeetCode, AtCoder, and CodeForces, evaluating code generation, self-repair, and execution.

MATH-500 Benchmark

A 500-problem subset from the MATH dataset, featuring competition-level mathematics across six domains including algebra, geometry, and number theory.

AIME 2025 Benchmark

All 30 problems from the 2025 American Invitational Mathematics Examination, testing olympiad-level mathematical reasoning with integer answers from 000-999.

Global-MMLU-Lite Benchmark

A lightweight, multilingual version of MMLU, designed to evaluate knowledge and reasoning skills across a diverse range of languages and cultural contexts.

Artificial Analysis Openness Index

背景

方法论

相关链接

Openness Index

Openness Index

Openness Index

Artificial Analysis Openness Index：得分

指数组成

Artificial Analysis Openness Index：组成部分

透明度与可用性

Artificial Analysis Openness Index：模型可用性与透明度

Openness Index 与发布日期

Artificial Analysis Openness Index：得分与发布日期

开放性与智能水平

Artificial Analysis Openness Index 与 Artificial Analysis Intelligence Index

Openness Index 的构成

评分方法

Openness Index 排行榜

常见问题

什么是 Artificial Analysis Openness Index？

Artificial Analysis Openness Index 如何计算？

哪个 AI 模型的 Artificial Analysis Openness Index 分数最高？

为什么 Artificial Analysis Openness Index 很重要？

开放权重等同于开源吗？

探索评测